language-servers/roc
1
0
Fork 0
mirror of https://github.com/roc-lang/roc.git synced 2025-10-02 16:21:11 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 6

Merge branch 'trunk' of github.com:rtfeldman/roc into single-quote-literal

Browse source
This commit is contained in:
Eric Correia 2021-10-02 09:55:58 -04:00
commit 555478cdf0
108 changed files with 7511 additions and 1611 deletions
Download patch file Download diff file Expand all files Collapse all files

1
.github/workflows/benchmarks.yml vendored
View file

@ -7,6 +7,7 @@ name: Benchmarks
env:
RUST_BACKTRACE: 1
ROC_NUM_WORKERS: 1
jobs:
prep-dependency-container:

1
.gitignore vendored
View file

@ -3,6 +3,7 @@ generated-docs
zig-cache
.direnv
*.rs.bk
*.o
# llvm human-readable output
*.ll

3
BUILDING_FROM_SOURCE.md
View file

@ -7,6 +7,7 @@ To build the compiler, you need these installed:
* Python 2.7 (Windows only), `python-is-python3` (Ubuntu)
* [Zig](https://ziglang.org/), see below for version
* `libxkbcommon` - macOS seems to have it already; on Ubuntu or Debian you can get it with `apt-get install libxkbcommon-dev`
* LLVM, see below for version
To run the test suite (via `cargo test`), you additionally need to install:
@ -73,6 +74,8 @@ There are also alternative installation options at http://releases.llvm.org/down
## Using Nix
:exclamation: **Our Nix setup is currently broken, you'll have to install manually for now** :exclamation:
### Install
Using [nix](https://nixos.org/download.html) is a quick way to get an environment bootstrapped with a single command.

43
Cargo.lock generated
View file

@ -1720,6 +1720,16 @@ version = "2.1.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9a3a5bfb195931eeb336b2a7b4d761daec841b97f947d34394601737a7bba5e4"
[[package]]
name = "iced-x86"
version = "1.14.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "7383772b06135cede839b7270023b46403656a9148024886e721e82639d3f90e"
dependencies = [
"lazy_static",
"static_assertions",
]
[[package]]
name = "ident_case"
version = "1.0.1"
@ -2147,6 +2157,15 @@ dependencies = [
"libc",
]
[[package]]
name = "memmap2"
version = "0.3.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "00b6c2ebff6180198788f5db08d7ce3bc1d0b617176678831a7510825973e357"
dependencies = [
"libc",
]
[[package]]
name = "memoffset"
version = "0.5.6"
@ -2498,6 +2517,9 @@ version = "0.26.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "39f37e50073ccad23b6d09bcb5b263f4e76d3bb6038e4a3c08e52162ffa8abc2"
dependencies = [
"crc32fast",
"flate2",
"indexmap",
"memchr",
]
@ -3380,7 +3402,9 @@ dependencies = [
"roc_can",
"roc_collections",
"roc_constrain",
"roc_gen_dev",
"roc_gen_llvm",
"roc_gen_wasm",
"roc_load",
"roc_module",
"roc_mono",
@ -3463,6 +3487,7 @@ dependencies = [
"roc_editor",
"roc_fmt",
"roc_gen_llvm",
"roc_linker",
"roc_load",
"roc_module",
"roc_mono",
@ -3695,6 +3720,24 @@ dependencies = [
name = "roc_ident"
version = "0.1.0"
[[package]]
name = "roc_linker"
version = "0.1.0"
dependencies = [
"bincode",
"bumpalo",
"clap 3.0.0-beta.1",
"iced-x86",
"memmap2 0.3.1",
"object 0.26.2",
"roc_build",
"roc_collections",
"roc_mono",
"serde",
"target-lexicon",
"tempfile",
]
[[package]]
name = "roc_load"
version = "0.1.0"

1
Cargo.toml
View file

@ -33,6 +33,7 @@ members = [
"cli/cli_utils",
"roc_std",
"docs",
"linker",
]
exclude = [ "ci/bench-runner" ]
# Needed to be able to run `cargo run -p roc_cli --no-default-features` -

4
Earthfile
View file

@ -46,7 +46,7 @@ install-zig-llvm-valgrind-clippy-rustfmt:
copy-dirs:
FROM +install-zig-llvm-valgrind-clippy-rustfmt
COPY --dir cli compiler docs editor roc_std vendor examples Cargo.toml Cargo.lock ./
COPY --dir cli compiler docs editor roc_std vendor examples linker Cargo.toml Cargo.lock ./
test-zig:
FROM +install-zig-llvm-valgrind-clippy-rustfmt
@ -66,7 +66,7 @@ check-rustfmt:
check-typos:
RUN cargo install --version 1.0.11 typos-cli
COPY --dir .github ci cli compiler docs editor examples nightly_benches packages roc_std www *.md LEGAL_DETAILS shell.nix ./
COPY --dir .github ci cli compiler docs editor examples linker nightly_benches packages roc_std www *.md LEGAL_DETAILS shell.nix ./
RUN typos
test-rust:

1
cli/Cargo.toml
View file

@ -57,6 +57,7 @@ roc_build = { path = "../compiler/build", default-features = false }
roc_fmt = { path = "../compiler/fmt" }
roc_reporting = { path = "../compiler/reporting" }
roc_editor = { path = "../editor", optional = true }
roc_linker = { path = "../linker" }
# TODO switch to clap 3.0.0 once it's out. Tried adding clap = "~3.0.0-beta.1" and cargo wouldn't accept it
clap = { git = "https://github.com/rtfeldman/clap", branch = "master" }
const_format = "0.2"

294
cli/src/build.rs
View file

@ -3,6 +3,7 @@ use roc_build::{
link::{link, rebuild_host, LinkType},
program,
};
use roc_builtins::bitcode;
use roc_can::builtins::builtin_defs_map;
use roc_collections::all::MutMap;
use roc_load::file::LoadingProblem;
@ -43,6 +44,7 @@ pub struct BuiltFile {
}
#[cfg(feature = "llvm")]
#[allow(clippy::too_many_arguments)]
pub fn build_file<'a>(
arena: &'a Bump,
target: &Triple,
@ -50,7 +52,10 @@ pub fn build_file<'a>(
roc_file_path: PathBuf,
opt_level: OptLevel,
emit_debug_info: bool,
emit_timings: bool,
link_type: LinkType,
surgically_link: bool,
precompiled: bool,
) -> Result<BuiltFile, LoadingProblem<'a>> {
let compilation_start = SystemTime::now();
let ptr_bytes = target.pointer_width().unwrap().bytes() as u32;
@ -59,10 +64,7 @@ pub fn build_file<'a>(
let subs_by_module = MutMap::default();
// Release builds use uniqueness optimizations
let stdlib = match opt_level {
OptLevel::Normal => arena.alloc(roc_builtins::std::standard_stdlib()),
OptLevel::Optimize => arena.alloc(roc_builtins::std::standard_stdlib()),
};
let stdlib = arena.alloc(roc_builtins::std::standard_stdlib());
let loaded = roc_load::file::load_and_monomorphize(
arena,
@ -75,16 +77,7 @@ pub fn build_file<'a>(
)?;
use target_lexicon::Architecture;
let emit_wasm = match target.architecture {
Architecture::X86_64 => false,
Architecture::Aarch64(_) => false,
Architecture::Wasm32 => true,
Architecture::X86_32(_) => false,
_ => panic!(
"TODO gracefully handle unsupported architecture: {:?}",
target.architecture
),
};
let emit_wasm = matches!(target.architecture, Architecture::Wasm32);
// TODO wasm host extension should be something else ideally
// .bc does not seem to work because
@ -95,7 +88,39 @@ pub fn build_file<'a>(
let host_extension = if emit_wasm { "zig" } else { "o" };
let app_extension = if emit_wasm { "bc" } else { "o" };
let cwd = roc_file_path.parent().unwrap();
let mut binary_path = cwd.join(&*loaded.output_path); // TODO should join ".exe" on Windows
if emit_wasm {
binary_path.set_extension("wasm");
}
let mut host_input_path = PathBuf::from(cwd);
let path_to_platform = loaded.platform_path.clone();
host_input_path.push(&*path_to_platform);
host_input_path.push("host");
host_input_path.set_extension(host_extension);
// TODO this should probably be moved before load_and_monomorphize.
// To do this we will need to preprocess files just for their exported symbols.
// Also, we should no longer need to do this once we have platforms on
// a package repository, as we can then get precompiled hosts from there.
let rebuild_thread = spawn_rebuild_thread(
opt_level,
surgically_link,
precompiled,
host_input_path.clone(),
binary_path.clone(),
target,
loaded
.exposed_to_host
.keys()
.map(|x| x.as_str(&loaded.interns).to_string())
.collect(),
);
// TODO try to move as much of this linking as possible to the precompiled
// host, to minimize the amount of host-application linking required.
let app_o_file = Builder::new()
.prefix("roc_app")
.suffix(&format!(".{}", app_extension))
@ -146,17 +171,26 @@ pub fn build_file<'a>(
}
}
let cwd = roc_file_path.parent().unwrap();
let binary_path = cwd.join(&*loaded.output_path); // TODO should join ".exe" on Windows
let code_gen_timing = program::gen_from_mono_module(
arena,
loaded,
&roc_file_path,
target,
app_o_file,
opt_level,
emit_debug_info,
);
// This only needs to be mutable for report_problems. This can't be done
// inside a nested scope without causing a borrow error!
let mut loaded = loaded;
program::report_problems(&mut loaded);
let loaded = loaded;
let code_gen_timing = match opt_level {
OptLevel::Normal | OptLevel::Optimize => program::gen_from_mono_module_llvm(
arena,
loaded,
&roc_file_path,
target,
app_o_file,
opt_level,
emit_debug_info,
),
OptLevel::Development => {
program::gen_from_mono_module_dev(arena, loaded, target, app_o_file)
}
};
buf.push('\n');
buf.push_str(" ");
@ -177,7 +211,7 @@ pub fn build_file<'a>(
})
.len();
if emit_debug_info {
if emit_timings {
println!(
"\n\nCompilation finished!\n\nHere's how long each module took to compile:\n\n{}",
buf
@ -190,65 +224,189 @@ pub fn build_file<'a>(
);
}
// Step 2: link the precompiled host and compiled app
let mut host_input_path = PathBuf::from(cwd);
host_input_path.push(&*path_to_platform);
host_input_path.push("host");
host_input_path.set_extension(host_extension);
// TODO we should no longer need to do this once we have platforms on
// a package repository, as we can then get precompiled hosts from there.
let rebuild_host_start = SystemTime::now();
rebuild_host(target, host_input_path.as_path());
let rebuild_host_end = rebuild_host_start.elapsed().unwrap();
if emit_debug_info {
let rebuild_duration = rebuild_thread.join().unwrap();
if emit_timings && !precompiled {
println!(
"Finished rebuilding the host in {} ms\n",
rebuild_host_end.as_millis()
"Finished rebuilding and preprocessing the host in {} ms\n",
rebuild_duration
);
}
// TODO try to move as much of this linking as possible to the precompiled
// host, to minimize the amount of host-application linking required.
// Step 2: link the precompiled host and compiled app
let link_start = SystemTime::now();
let (mut child, binary_path) = // TODO use lld
link(
target,
binary_path,
&[host_input_path.as_path().to_str().unwrap(), app_o_file.to_str().unwrap()],
link_type
)
.map_err(|_| {
todo!("gracefully handle `rustc` failing to spawn.");
let outcome = if surgically_link {
roc_linker::link_preprocessed_host(target, &host_input_path, app_o_file, &binary_path)
.map_err(|_| {
todo!("gracefully handle failing to surgically link");
})?;
BuildOutcome::NoProblems
} else {
let mut inputs = vec![
host_input_path.as_path().to_str().unwrap(),
app_o_file.to_str().unwrap(),
];
if matches!(opt_level, OptLevel::Development) {
inputs.push(bitcode::OBJ_PATH);
}
let (mut child, _) = // TODO use lld
link(
target,
binary_path.clone(),
&inputs,
link_type
)
.map_err(|_| {
todo!("gracefully handle `ld` failing to spawn.");
})?;
let exit_status = child.wait().map_err(|_| {
todo!("gracefully handle error after `ld` spawned");
})?;
let cmd_result = child.wait().map_err(|_| {
todo!("gracefully handle error after `rustc` spawned");
});
// TODO change this to report whether there were errors or warnings!
if exit_status.success() {
BuildOutcome::NoProblems
} else {
BuildOutcome::Errors
}
};
let linking_time = link_start.elapsed().unwrap();
if emit_debug_info {
if emit_timings {
println!("Finished linking in {} ms\n", linking_time.as_millis());
}
let total_time = compilation_start.elapsed().unwrap();
// If the cmd errored out, return the Err.
let exit_status = cmd_result?;
// TODO change this to report whether there were errors or warnings!
let outcome = if exit_status.success() {
BuildOutcome::NoProblems
} else {
BuildOutcome::Errors
};
Ok(BuiltFile {
binary_path,
outcome,
total_time,
})
}
fn spawn_rebuild_thread(
opt_level: OptLevel,
surgically_link: bool,
precompiled: bool,
host_input_path: PathBuf,
binary_path: PathBuf,
target: &Triple,
exported_symbols: Vec<String>,
) -> std::thread::JoinHandle<u128> {
let thread_local_target = target.clone();
std::thread::spawn(move || {
let rebuild_host_start = SystemTime::now();
if !precompiled {
if surgically_link {
roc_linker::build_and_preprocess_host(
opt_level,
&thread_local_target,
host_input_path.as_path(),
exported_symbols,
)
.unwrap();
} else {
rebuild_host(
opt_level,
&thread_local_target,
host_input_path.as_path(),
None,
);
}
}
if surgically_link {
// Copy preprocessed host to executable location.
let prehost = host_input_path.with_file_name("preprocessedhost");
std::fs::copy(prehost, binary_path.as_path()).unwrap();
}
let rebuild_host_end = rebuild_host_start.elapsed().unwrap();
rebuild_host_end.as_millis()
})
}
#[allow(clippy::too_many_arguments)]
pub fn check_file(
arena: &Bump,
src_dir: PathBuf,
roc_file_path: PathBuf,
emit_timings: bool,
) -> Result<usize, LoadingProblem> {
let compilation_start = SystemTime::now();
// only used for generating errors. We don't do code generation, so hardcoding should be fine
// we need monomorphization for when exhaustiveness checking
let ptr_bytes = 8;
// Step 1: compile the app and generate the .o file
let subs_by_module = MutMap::default();
// Release builds use uniqueness optimizations
let stdlib = arena.alloc(roc_builtins::std::standard_stdlib());
let mut loaded = roc_load::file::load_and_monomorphize(
arena,
roc_file_path,
stdlib,
src_dir.as_path(),
subs_by_module,
ptr_bytes,
builtin_defs_map,
)?;
let buf = &mut String::with_capacity(1024);
let mut it = loaded.timings.iter().peekable();
while let Some((module_id, module_timing)) = it.next() {
let module_name = loaded.interns.module_name(*module_id);
buf.push_str(" ");
if module_name.is_empty() {
// the App module
buf.push_str("Application Module");
} else {
buf.push_str(module_name);
}
buf.push('\n');
report_timing(buf, "Read .roc file from disk", module_timing.read_roc_file);
report_timing(buf, "Parse header", module_timing.parse_header);
report_timing(buf, "Parse body", module_timing.parse_body);
report_timing(buf, "Canonicalize", module_timing.canonicalize);
report_timing(buf, "Constrain", module_timing.constrain);
report_timing(buf, "Solve", module_timing.solve);
report_timing(
buf,
"Find Specializations",
module_timing.find_specializations,
);
report_timing(
buf,
"Make Specializations",
module_timing.make_specializations,
);
report_timing(buf, "Other", module_timing.other());
buf.push('\n');
report_timing(buf, "Total", module_timing.total());
if it.peek().is_some() {
buf.push('\n');
}
}
let compilation_end = compilation_start.elapsed().unwrap();
if emit_timings {
println!(
"\n\nCompilation finished!\n\nHere's how long each module took to compile:\n\n{}",
buf
);
println!("Finished checking in {} ms\n", compilation_end.as_millis(),);
}
Ok(program::report_problems(&mut loaded))
}

108
cli/src/lib.rs
View file

@ -26,11 +26,16 @@ pub const CMD_BUILD: &str = "build";
pub const CMD_REPL: &str = "repl";
pub const CMD_EDIT: &str = "edit";
pub const CMD_DOCS: &str = "docs";
pub const CMD_CHECK: &str = "check";
pub const FLAG_DEBUG: &str = "debug";
pub const FLAG_DEV: &str = "dev";
pub const FLAG_OPTIMIZE: &str = "optimize";
pub const FLAG_LIB: &str = "lib";
pub const FLAG_BACKEND: &str = "backend";
pub const FLAG_TIME: &str = "time";
pub const FLAG_LINK: &str = "roc-linker";
pub const FLAG_PRECOMPILED: &str = "precompiled-host";
pub const ROC_FILE: &str = "ROC_FILE";
pub const BACKEND: &str = "BACKEND";
pub const DIRECTORY_OR_FILES: &str = "DIRECTORY_OR_FILES";
@ -53,6 +58,12 @@ pub fn build_app<'a>() -> App<'a> {
.help("Optimize your compiled Roc program to run faster. (Optimization takes time to complete.)")
.required(false),
)
.arg(
Arg::with_name(FLAG_DEV)
.long(FLAG_DEV)
.help("Make compilation as fast as possible. (Runtime performance may suffer)")
.required(false),
)
.arg(
Arg::with_name(FLAG_BACKEND)
.long(FLAG_BACKEND)
@ -74,6 +85,24 @@ pub fn build_app<'a>() -> App<'a> {
.help("Store LLVM debug information in the generated program")
.required(false),
)
.arg(
Arg::with_name(FLAG_TIME)
.long(FLAG_TIME)
.help("Prints detailed compilation time information.")
.required(false),
)
.arg(
Arg::with_name(FLAG_LINK)
.long(FLAG_LINK)
.help("Uses the roc linker instead of the system linker.")
.required(false),
)
.arg(
Arg::with_name(FLAG_PRECOMPILED)
.long(FLAG_PRECOMPILED)
.help("Assumes the host has been precompiled and skips recompiling the host.")
.required(false),
)
)
.subcommand(App::new(CMD_RUN)
.about("DEPRECATED - now use `roc [FILE]` instead of `roc run [FILE]`")
@ -84,6 +113,12 @@ pub fn build_app<'a>() -> App<'a> {
.help("Optimize the compiled program to run faster. (Optimization takes time to complete.)")
.required(false),
)
.arg(
Arg::with_name(FLAG_DEV)
.long(FLAG_DEV)
.help("Make compilation as fast as possible. (Runtime performance may suffer)")
.required(false),
)
.arg(
Arg::with_name(FLAG_DEBUG)
.long(FLAG_DEBUG)
@ -104,6 +139,20 @@ pub fn build_app<'a>() -> App<'a> {
.subcommand(App::new(CMD_REPL)
.about("Launch the interactive Read Eval Print Loop (REPL)")
)
.subcommand(App::new(CMD_CHECK)
.about("Build a binary from the given .roc file, but don't run it")
.arg(
Arg::with_name(FLAG_TIME)
.long(FLAG_TIME)
.help("Prints detailed compilation time information.")
.required(false),
)
.arg(
Arg::with_name(ROC_FILE)
.help("The .roc file of an app to run")
.required(true),
)
)
.subcommand(
App::new(CMD_DOCS)
.about("Generate documentation for Roc modules")
@ -123,6 +172,12 @@ pub fn build_app<'a>() -> App<'a> {
.requires(ROC_FILE)
.required(false),
)
.arg(
Arg::with_name(FLAG_DEV)
.long(FLAG_DEV)
.help("Make compilation as fast as possible. (Runtime performance may suffer)")
.required(false),
)
.arg(
Arg::with_name(FLAG_DEBUG)
.long(FLAG_DEBUG)
@ -130,6 +185,24 @@ pub fn build_app<'a>() -> App<'a> {
.requires(ROC_FILE)
.required(false),
)
.arg(
Arg::with_name(FLAG_TIME)
.long(FLAG_TIME)
.help("Prints detailed compilation time information.")
.required(false),
)
.arg(
Arg::with_name(FLAG_LINK)
.long(FLAG_LINK)
.help("Uses the roc linker instead of the system linker.")
.required(false),
)
.arg(
Arg::with_name(FLAG_PRECOMPILED)
.long(FLAG_PRECOMPILED)
.help("Assumes the host has been precompiled and skips recompiling the host.")
.required(false),
)
.arg(
Arg::with_name(FLAG_BACKEND)
.long(FLAG_BACKEND)
@ -197,18 +270,31 @@ pub fn build(matches: &ArgMatches, config: BuildConfig) -> io::Result<i32> {
let filename = matches.value_of(ROC_FILE).unwrap();
let original_cwd = std::env::current_dir()?;
let opt_level = if matches.is_present(FLAG_OPTIMIZE) {
OptLevel::Optimize
} else {
OptLevel::Normal
let opt_level = match (
matches.is_present(FLAG_OPTIMIZE),
matches.is_present(FLAG_DEV),
) {
(true, false) => OptLevel::Optimize,
(true, true) => panic!("development cannot be optimized!"),
(false, true) => OptLevel::Development,
(false, false) => OptLevel::Normal,
};
let emit_debug_info = matches.is_present(FLAG_DEBUG);
let emit_timings = matches.is_present(FLAG_TIME);
let link_type = if matches.is_present(FLAG_LIB) {
LinkType::Dylib
} else {
LinkType::Executable
};
let surgically_link = matches.is_present(FLAG_LINK);
let precompiled = matches.is_present(FLAG_PRECOMPILED);
if surgically_link && !roc_linker::supported(&link_type, &target) {
panic!(
"Link type, {:?}, with target, {}, not supported by roc linker",
link_type, target
);
}
let path = Path::new(filename);
@ -239,7 +325,10 @@ pub fn build(matches: &ArgMatches, config: BuildConfig) -> io::Result<i32> {
path,
opt_level,
emit_debug_info,
emit_timings,
link_type,
surgically_link,
precompiled,
);
match res_binary_path {
@ -393,9 +482,7 @@ enum Backend {
Host,
X86_32,
X86_64,
Dev,
Wasm32,
Wasm32Dev,
}
impl Default for Backend {
@ -410,9 +497,7 @@ impl Backend {
Backend::Host => "host",
Backend::X86_32 => "x86_32",
Backend::X86_64 => "x86_64",
Backend::Dev => "dev",
Backend::Wasm32 => "wasm32",
Backend::Wasm32Dev => "wasm32_dev",
}
}
@ -421,9 +506,7 @@ impl Backend {
Backend::Host.as_str(),
Backend::X86_32.as_str(),
Backend::X86_64.as_str(),
Backend::Dev.as_str(),
Backend::Wasm32.as_str(),
Backend::Wasm32Dev.as_str(),
];
fn to_triple(&self) -> Triple {
@ -446,8 +529,7 @@ impl Backend {
triple
}
Backend::Dev => todo!(),
Backend::Wasm32 | Backend::Wasm32Dev => {
Backend::Wasm32 => {
triple.architecture = Architecture::Wasm32;
triple.binary_format = BinaryFormat::Wasm;
@ -471,9 +553,7 @@ impl std::str::FromStr for Backend {
"host" => Ok(Backend::Host),
"x86_32" => Ok(Backend::X86_32),
"x86_64" => Ok(Backend::X86_64),
"dev" => Ok(Backend::Dev),
"wasm32" => Ok(Backend::Wasm32),
"wasm32_dev" => Ok(Backend::Wasm32Dev),
_ => Err(()),
}
}

31
cli/src/main.rs
View file

@ -1,7 +1,9 @@
use roc_cli::build::check_file;
use roc_cli::{
build_app, docs, repl, BuildConfig, CMD_BUILD, CMD_DOCS, CMD_EDIT, CMD_REPL, CMD_RUN,
DIRECTORY_OR_FILES, ROC_FILE,
build_app, docs, repl, BuildConfig, CMD_BUILD, CMD_CHECK, CMD_DOCS, CMD_EDIT, CMD_REPL,
CMD_RUN, DIRECTORY_OR_FILES, FLAG_TIME, ROC_FILE,
};
use roc_load::file::LoadingProblem;
use std::fs::{self, FileType};
use std::io;
use std::path::{Path, PathBuf};
@ -52,6 +54,31 @@ If you're building the compiler from source you'll want to do `cargo run [FILE]`
Ok(1)
}
Some(CMD_CHECK) => {
let arena = bumpalo::Bump::new();
let matches = matches.subcommand_matches(CMD_CHECK).unwrap();
let emit_timings = matches.is_present(FLAG_TIME);
let filename = matches.value_of(ROC_FILE).unwrap();
let roc_file_path = PathBuf::from(filename);
let src_dir = roc_file_path.parent().unwrap().to_owned();
match check_file(&arena, src_dir, roc_file_path, emit_timings) {
Ok(number_of_errors) => {
let exit_code = if number_of_errors != 0 { 1 } else { 0 };
Ok(exit_code)
}
Err(LoadingProblem::FormattedReport(report)) => {
print!("{}", report);
Ok(1)
}
Err(other) => {
panic!("build_file failed with error:\n{:?}", other);
}
}
}
Some(CMD_REPL) => {
repl::main()?;

7
cli/src/repl/eval.rs
View file

@ -353,6 +353,13 @@ fn jit_to_ast_help<'a>(
| Layout::RecursivePointer => {
todo!("add support for rendering recursive tag unions in the REPL")
}
Layout::LambdaSet(lambda_set) => jit_to_ast_help(
env,
lib,
main_fn_name,
&lambda_set.runtime_representation(),
content,
),
}
}

9
cli/src/repl/gen.rs
View file

@ -107,12 +107,13 @@ pub fn gen_and_eval<'a>(
}
for problem in type_problems {
let report = type_problem(&alloc, module_path.clone(), problem);
let mut buf = String::new();
if let Some(report) = type_problem(&alloc, module_path.clone(), problem) {
let mut buf = String::new();
report.render_color_terminal(&mut buf, &alloc, &palette);
report.render_color_terminal(&mut buf, &alloc, &palette);
lines.push(buf);
lines.push(buf);
}
}
for problem in mono_problems {

25
cli/tests/cli_run.rs
View file

@ -157,6 +157,15 @@ mod cli_run {
let example = $example;
let file_name = example_file(dir_name, example.filename);
match example.executable_filename {
"hello-web" => {
// this is a web webassembly example, but we don't test with JS at the moment
eprintln!("WARNING: skipping testing example {} because the test is broken right now!", example.filename);
return;
}
_ => {}
}
// Check with and without optimizations
check_output_with_stdin(
&file_name,
@ -224,6 +233,20 @@ mod cli_run {
expected_ending:"Hello, World!\n",
use_valgrind: true,
},
hello_web:"hello-web" => Example {
filename: "Hello.roc",
executable_filename: "hello-web",
stdin: &[],
expected_ending:"Hello, World!\n",
use_valgrind: true,
},
fib:"fib" => Example {
filename: "Fib.roc",
executable_filename: "fib",
stdin: &[],
expected_ending:"55\n",
use_valgrind: true,
},
quicksort:"quicksort" => Example {
filename: "Quicksort.roc",
executable_filename: "quicksort",
@ -242,7 +265,7 @@ mod cli_run {
filename: "Main.roc",
executable_filename: "effect-example",
stdin: &["hi there!"],
expected_ending: "hi there!\n",
expected_ending: "hi there!\nIt is known\n",
use_valgrind: true,
},
// tea:"tea" => Example {

14
cli/tests/fixtures/multi-dep-str/platform/host.zig vendored
View file

@ -22,7 +22,8 @@ comptime {
const mem = std.mem;
const Allocator = mem.Allocator;
extern fn roc__mainForHost_1_exposed(*RocCallResult) void;
extern fn roc__mainForHost_1_exposed() RocStr;
extern fn malloc(size: usize) callconv(.C) ?*c_void;
extern fn realloc(c_ptr: [*]align(@alignOf(u128)) u8, size: usize) callconv(.C) ?*c_void;
@ -47,28 +48,23 @@ export fn roc_panic(c_ptr: *c_void, tag_id: u32) callconv(.C) void {
std.process.exit(0);
}
const RocCallResult = extern struct { flag: usize, content: RocStr };
const Unit = extern struct {};
pub export fn main() i32 {
const stdout = std.io.getStdOut().writer();
const stderr = std.io.getStdErr().writer();
// make space for the result
var callresult = RocCallResult{ .flag = 0, .content = RocStr.empty() };
// start time
var ts1: std.os.timespec = undefined;
std.os.clock_gettime(std.os.CLOCK_REALTIME, &ts1) catch unreachable;
// actually call roc to populate the callresult
roc__mainForHost_1_exposed(&callresult);
const callresult = roc__mainForHost_1_exposed();
// stdout the result
stdout.print("{s}\n", .{callresult.content.asSlice()}) catch unreachable;
stdout.print("{s}\n", .{callresult.asSlice()}) catch unreachable;
callresult.content.deinit();
callresult.deinit();
// end time
var ts2: std.os.timespec = undefined;

13
cli/tests/fixtures/multi-dep-thunk/platform/host.zig vendored
View file

@ -22,7 +22,7 @@ comptime {
const mem = std.mem;
const Allocator = mem.Allocator;
extern fn roc__mainForHost_1_exposed(*RocCallResult) void;
extern fn roc__mainForHost_1_exposed() RocStr;
extern fn malloc(size: usize) callconv(.C) ?*c_void;
extern fn realloc(c_ptr: [*]align(@alignOf(u128)) u8, size: usize) callconv(.C) ?*c_void;
@ -47,28 +47,23 @@ export fn roc_panic(c_ptr: *c_void, tag_id: u32) callconv(.C) void {
std.process.exit(0);
}
const RocCallResult = extern struct { flag: usize, content: RocStr };
const Unit = extern struct {};
pub export fn main() i32 {
const stdout = std.io.getStdOut().writer();
const stderr = std.io.getStdErr().writer();
// make space for the result
var callresult = RocCallResult{ .flag = 0, .content = RocStr.empty() };
// start time
var ts1: std.os.timespec = undefined;
std.os.clock_gettime(std.os.CLOCK_REALTIME, &ts1) catch unreachable;
// actually call roc to populate the callresult
roc__mainForHost_1_exposed(&callresult);
const callresult = roc__mainForHost_1_exposed();
// stdout the result
stdout.print("{s}\n", .{callresult.content.asSlice()}) catch unreachable;
stdout.print("{s}\n", .{callresult.asSlice()}) catch unreachable;
callresult.content.deinit();
callresult.deinit();
// end time
var ts2: std.os.timespec = undefined;

2
compiler/build/Cargo.toml
View file

@ -20,6 +20,8 @@ roc_solve = { path = "../solve" }
roc_mono = { path = "../mono" }
roc_load = { path = "../load" }
roc_gen_llvm = { path = "../gen_llvm", optional = true }
roc_gen_wasm = { path = "../gen_wasm" }
roc_gen_dev = { path = "../gen_dev" }
roc_reporting = { path = "../reporting" }
roc_std = { path = "../../roc_std" }
im = "14" # im and im-rc should always have the same version!

384
compiler/build/src/link.rs
View file

@ -1,6 +1,7 @@
use crate::target::arch_str;
#[cfg(feature = "llvm")]
use libloading::{Error, Library};
use roc_builtins::bitcode;
#[cfg(feature = "llvm")]
use roc_mono::ir::OptLevel;
use std::collections::HashMap;
@ -56,7 +57,7 @@ fn find_zig_str_path() -> PathBuf {
return zig_str_path;
}
panic!("cannot find `str.zig`")
panic!("cannot find `str.zig`. Launch me from either the root of the roc repo or one level down(roc/examples, roc/cli...)")
}
fn find_wasi_libc_path() -> PathBuf {
@ -76,6 +77,7 @@ fn find_wasi_libc_path() -> PathBuf {
}
#[cfg(not(target_os = "macos"))]
#[allow(clippy::too_many_arguments)]
pub fn build_zig_host_native(
env_path: &str,
env_home: &str,
@ -83,33 +85,48 @@ pub fn build_zig_host_native(
zig_host_src: &str,
zig_str_path: &str,
target: &str,
opt_level: OptLevel,
shared_lib_path: Option<&Path>,
) -> Output {
Command::new("zig")
let mut command = Command::new("zig");
command
.env_clear()
.env("PATH", env_path)
.env("HOME", env_home)
.args(&[
"build-obj",
zig_host_src,
emit_bin,
"--pkg-begin",
"str",
zig_str_path,
"--pkg-end",
// include the zig runtime
"-fcompiler-rt",
// include libc
"--library",
"c",
// cross-compile?
"-target",
target,
])
.output()
.unwrap()
.env("HOME", env_home);
if let Some(shared_lib_path) = shared_lib_path {
command.args(&[
"build-exe",
"-fPIE",
shared_lib_path.to_str().unwrap(),
bitcode::OBJ_PATH,
]);
} else {
command.args(&["build-obj", "-fPIC"]);
}
command.args(&[
zig_host_src,
emit_bin,
"--pkg-begin",
"str",
zig_str_path,
"--pkg-end",
// include the zig runtime
"-fcompiler-rt",
// include libc
"--library",
"c",
// cross-compile?
"-target",
target,
]);
if matches!(opt_level, OptLevel::Optimize) {
command.args(&["-O", "ReleaseSafe"]);
}
command.output().unwrap()
}
#[cfg(target_os = "macos")]
#[allow(clippy::too_many_arguments)]
pub fn build_zig_host_native(
env_path: &str,
env_home: &str,
@ -117,6 +134,8 @@ pub fn build_zig_host_native(
zig_host_src: &str,
zig_str_path: &str,
_target: &str,
opt_level: OptLevel,
shared_lib_path: Option<&Path>,
) -> Output {
use serde_json::Value;
@ -158,29 +177,41 @@ pub fn build_zig_host_native(
zig_compiler_rt_path.push("special");
zig_compiler_rt_path.push("compiler_rt.zig");
Command::new("zig")
let mut command = Command::new("zig");
command
.env_clear()
.env("PATH", &env_path)
.env("HOME", &env_home)
.args(&[
"build-obj",
zig_host_src,
emit_bin,
"--pkg-begin",
"str",
zig_str_path,
"--pkg-end",
// include the zig runtime
"--pkg-begin",
"compiler_rt",
zig_compiler_rt_path.to_str().unwrap(),
"--pkg-end",
// include libc
"--library",
"c",
])
.output()
.unwrap()
.env("HOME", &env_home);
if let Some(shared_lib_path) = shared_lib_path {
command.args(&[
"build-exe",
"-fPIE",
shared_lib_path.to_str().unwrap(),
bitcode::OBJ_PATH,
]);
} else {
command.args(&["build-obj", "-fPIC"]);
}
command.args(&[
zig_host_src,
emit_bin,
"--pkg-begin",
"str",
zig_str_path,
"--pkg-end",
// include the zig runtime
"--pkg-begin",
"compiler_rt",
zig_compiler_rt_path.to_str().unwrap(),
"--pkg-end",
// include libc
"--library",
"c",
]);
if matches!(opt_level, OptLevel::Optimize) {
command.args(&["-O", "ReleaseSafe"]);
}
command.output().unwrap()
}
pub fn build_zig_host_wasm32(
@ -189,7 +220,12 @@ pub fn build_zig_host_wasm32(
emit_bin: &str,
zig_host_src: &str,
zig_str_path: &str,
opt_level: OptLevel,
shared_lib_path: Option<&Path>,
) -> Output {
if shared_lib_path.is_some() {
unimplemented!("Linking a shared library to wasm not yet implemented");
}
// NOTE currently just to get compiler warnings if the host code is invalid.
// the produced artifact is not used
//
@ -198,7 +234,8 @@ pub fn build_zig_host_wasm32(
// we'd like to compile with `-target wasm32-wasi` but that is blocked on
//
// https://github.com/ziglang/zig/issues/9414
Command::new("zig")
let mut command = Command::new("zig");
command
.env_clear()
.env("PATH", env_path)
.env("HOME", env_home)
@ -219,19 +256,68 @@ pub fn build_zig_host_wasm32(
"i386-linux-musl",
// "wasm32-wasi",
// "-femit-llvm-ir=/home/folkertdev/roc/roc/examples/benchmarks/platform/host.ll",
])
.output()
.unwrap()
"-fPIC",
"--strip",
]);
if matches!(opt_level, OptLevel::Optimize) {
command.args(&["-O", "ReleaseSafe"]);
}
command.output().unwrap()
}
pub fn rebuild_host(target: &Triple, host_input_path: &Path) {
pub fn build_c_host_native(
env_path: &str,
env_home: &str,
dest: &str,
sources: &[&str],
opt_level: OptLevel,
shared_lib_path: Option<&Path>,
) -> Output {
let mut command = Command::new("clang");
command
.env_clear()
.env("PATH", &env_path)
.env("HOME", &env_home)
.args(sources)
.args(&["-o", dest]);
if let Some(shared_lib_path) = shared_lib_path {
command.args(&[
shared_lib_path.to_str().unwrap(),
bitcode::OBJ_PATH,
"-fPIE",
"-pie",
"-lm",
"-lpthread",
"-ldl",
"-lrt",
"-lutil",
]);
} else {
command.args(&["-fPIC", "-c"]);
}
if matches!(opt_level, OptLevel::Optimize) {
command.arg("-O2");
}
command.output().unwrap()
}
pub fn rebuild_host(
opt_level: OptLevel,
target: &Triple,
host_input_path: &Path,
shared_lib_path: Option<&Path>,
) {
let c_host_src = host_input_path.with_file_name("host.c");
let c_host_dest = host_input_path.with_file_name("c_host.o");
let zig_host_src = host_input_path.with_file_name("host.zig");
let rust_host_src = host_input_path.with_file_name("host.rs");
let rust_host_dest = host_input_path.with_file_name("rust_host.o");
let cargo_host_src = host_input_path.with_file_name("Cargo.toml");
let host_dest_native = host_input_path.with_file_name("host.o");
let host_dest_native = host_input_path.with_file_name(if shared_lib_path.is_some() {
"dynhost"
} else {
"host.o"
});
let host_dest_wasm = host_input_path.with_file_name("host.bc");
let env_path = env::var("PATH").unwrap_or_else(|_| "".to_string());
@ -257,6 +343,8 @@ pub fn rebuild_host(target: &Triple, host_input_path: &Path) {
&emit_bin,
zig_host_src.to_str().unwrap(),
zig_str_path.to_str().unwrap(),
opt_level,
shared_lib_path,
)
}
Architecture::X86_64 => {
@ -268,6 +356,8 @@ pub fn rebuild_host(target: &Triple, host_input_path: &Path) {
zig_host_src.to_str().unwrap(),
zig_str_path.to_str().unwrap(),
"native",
opt_level,
shared_lib_path,
)
}
Architecture::X86_32(_) => {
@ -279,87 +369,142 @@ pub fn rebuild_host(target: &Triple, host_input_path: &Path) {
zig_host_src.to_str().unwrap(),
zig_str_path.to_str().unwrap(),
"i386-linux-musl",
opt_level,
shared_lib_path,
)
}
_ => panic!("Unsupported architecture {:?}", target.architecture),
};
validate_output("host.zig", "zig", output)
} else {
// Compile host.c
let output = Command::new("clang")
.env_clear()
.env("PATH", &env_path)
.args(&[
"-c",
c_host_src.to_str().unwrap(),
"-o",
c_host_dest.to_str().unwrap(),
])
.output()
.unwrap();
validate_output("host.c", "clang", output);
}
if cargo_host_src.exists() {
} else if cargo_host_src.exists() {
// Compile and link Cargo.toml, if it exists
let cargo_dir = host_input_path.parent().unwrap();
let libhost_dir = cargo_dir.join("target").join("release");
let libhost_dir =
cargo_dir
.join("target")
.join(if matches!(opt_level, OptLevel::Optimize) {
"release"
} else {
"debug"
});
let libhost = libhost_dir.join("libhost.a");
let output = Command::new("cargo")
.args(&["build", "--release"])
.current_dir(cargo_dir)
.output()
.unwrap();
let mut command = Command::new("cargo");
command.arg("build").current_dir(cargo_dir);
if matches!(opt_level, OptLevel::Optimize) {
command.arg("--release");
}
let output = command.output().unwrap();
validate_output("src/lib.rs", "cargo build --release", output);
validate_output("src/lib.rs", "cargo build", output);
let output = Command::new("ld")
.env_clear()
.env("PATH", &env_path)
.args(&[
"-r",
"-L",
libhost_dir.to_str().unwrap(),
c_host_dest.to_str().unwrap(),
"-lhost",
"-o",
// Cargo hosts depend on a c wrapper for the api. Compile host.c as well.
if shared_lib_path.is_some() {
// If compiling to executable, let c deal with linking as well.
let output = build_c_host_native(
&env_path,
&env_home,
host_dest_native.to_str().unwrap(),
])
.output()
.unwrap();
&[c_host_src.to_str().unwrap(), libhost.to_str().unwrap()],
opt_level,
shared_lib_path,
);
validate_output("host.c", "clang", output);
} else {
let output = build_c_host_native(
&env_path,
&env_home,
c_host_dest.to_str().unwrap(),
&[c_host_src.to_str().unwrap()],
opt_level,
shared_lib_path,
);
validate_output("host.c", "clang", output);
validate_output("c_host.o", "ld", output);
let output = Command::new("ld")
.env_clear()
.env("PATH", &env_path)
.args(&[
"-r",
"-L",
libhost_dir.to_str().unwrap(),
c_host_dest.to_str().unwrap(),
"-lhost",
"-o",
host_dest_native.to_str().unwrap(),
])
.output()
.unwrap();
validate_output("c_host.o", "ld", output);
// Clean up c_host.o
let output = Command::new("rm")
.env_clear()
.args(&["-f", c_host_dest.to_str().unwrap()])
.output()
.unwrap();
validate_output("rust_host.o", "rm", output);
}
} else if rust_host_src.exists() {
// Compile and link host.rs, if it exists
let output = Command::new("rustc")
.args(&[
rust_host_src.to_str().unwrap(),
"-o",
rust_host_dest.to_str().unwrap(),
])
.output()
.unwrap();
let mut command = Command::new("rustc");
command.args(&[
rust_host_src.to_str().unwrap(),
"-o",
rust_host_dest.to_str().unwrap(),
]);
if matches!(opt_level, OptLevel::Optimize) {
command.arg("-O");
}
let output = command.output().unwrap();
validate_output("host.rs", "rustc", output);
let output = Command::new("ld")
.env_clear()
.env("PATH", &env_path)
.args(&[
"-r",
c_host_dest.to_str().unwrap(),
rust_host_dest.to_str().unwrap(),
"-o",
// Rust hosts depend on a c wrapper for the api. Compile host.c as well.
if shared_lib_path.is_some() {
// If compiling to executable, let c deal with linking as well.
let output = build_c_host_native(
&env_path,
&env_home,
host_dest_native.to_str().unwrap(),
])
.output()
.unwrap();
&[
c_host_src.to_str().unwrap(),
rust_host_dest.to_str().unwrap(),
],
opt_level,
shared_lib_path,
);
validate_output("host.c", "clang", output);
} else {
let output = build_c_host_native(
&env_path,
&env_home,
c_host_dest.to_str().unwrap(),
&[c_host_src.to_str().unwrap()],
opt_level,
shared_lib_path,
);
validate_output("rust_host.o", "ld", output);
validate_output("host.c", "clang", output);
let output = Command::new("ld")
.env_clear()
.env("PATH", &env_path)
.args(&[
"-r",
c_host_dest.to_str().unwrap(),
rust_host_dest.to_str().unwrap(),
"-o",
host_dest_native.to_str().unwrap(),
])
.output()
.unwrap();
// Clean up rust_host.o
validate_output("rust_host.o", "ld", output);
}
// Clean up rust_host.o and c_host.o
let output = Command::new("rm")
.env_clear()
.args(&[
@ -371,15 +516,17 @@ pub fn rebuild_host(target: &Triple, host_input_path: &Path) {
.unwrap();
validate_output("rust_host.o", "rm", output);
} else if c_host_dest.exists() {
// Clean up c_host.o
let output = Command::new("mv")
.env_clear()
.args(&[c_host_dest, host_dest_native])
.output()
.unwrap();
validate_output("c_host.o", "mv", output);
} else if c_host_src.exists() {
// Compile host.c, if it exists
let output = build_c_host_native(
&env_path,
&env_home,
host_dest_native.to_str().unwrap(),
&[c_host_src.to_str().unwrap()],
opt_level,
shared_lib_path,
);
validate_output("host.c", "clang", output);
}
}
@ -521,6 +668,7 @@ fn link_linux(
"--eh-frame-hdr",
"-arch",
arch_str(target),
"-pie",
libcrt_path.join("crti.o").to_str().unwrap(),
libcrt_path.join("crtn.o").to_str().unwrap(),
])

177
compiler/build/src/program.rs
View file

@ -10,6 +10,8 @@ use roc_mono::ir::OptLevel;
use std::path::{Path, PathBuf};
use std::time::Duration;
use roc_collections::all::{MutMap, MutSet};
#[derive(Debug, Clone, Copy, Default)]
pub struct CodeGenTiming {
pub code_gen: Duration,
@ -20,33 +22,15 @@ pub struct CodeGenTiming {
// llvm we're using, consider moving me somewhere else.
const LLVM_VERSION: &str = "12";
// TODO how should imported modules factor into this? What if those use builtins too?
// TODO this should probably use more helper functions
// TODO make this polymorphic in the llvm functions so it can be reused for another backend.
#[cfg(feature = "llvm")]
#[allow(clippy::cognitive_complexity)]
pub fn gen_from_mono_module(
arena: &bumpalo::Bump,
mut loaded: MonomorphizedModule,
roc_file_path: &Path,
target: &target_lexicon::Triple,
app_o_file: &Path,
opt_level: OptLevel,
emit_debug_info: bool,
) -> CodeGenTiming {
use crate::target::{self, convert_opt_level};
use inkwell::attributes::{Attribute, AttributeLoc};
use inkwell::context::Context;
use inkwell::module::Linkage;
use inkwell::targets::{CodeModel, FileType, RelocMode};
use std::time::SystemTime;
// TODO instead of finding exhaustiveness problems in monomorphization, find
// them after type checking (like Elm does) so we can complete the entire
// `roc check` process without needing to monomorphize.
/// Returns the number of problems reported.
pub fn report_problems(loaded: &mut MonomorphizedModule) -> usize {
use roc_reporting::report::{
can_problem, mono_problem, type_problem, Report, RocDocAllocator, Severity::*,
DEFAULT_PALETTE,
};
let code_gen_start = SystemTime::now();
let palette = DEFAULT_PALETTE;
// This will often over-allocate total memory, but it means we definitely
@ -55,10 +39,10 @@ pub fn gen_from_mono_module(
let mut warnings = Vec::with_capacity(total_problems);
let mut errors = Vec::with_capacity(total_problems);
for (home, (module_path, src)) in loaded.sources {
for (home, (module_path, src)) in loaded.sources.iter() {
let mut src_lines: Vec<&str> = Vec::new();
if let Some((_, header_src)) = loaded.header_sources.get(&home) {
if let Some((_, header_src)) = loaded.header_sources.get(home) {
src_lines.extend(header_src.split('\n'));
src_lines.extend(src.split('\n').skip(1));
} else {
@ -66,9 +50,10 @@ pub fn gen_from_mono_module(
}
// Report parsing and canonicalization problems
let alloc = RocDocAllocator::new(&src_lines, home, &loaded.interns);
let alloc = RocDocAllocator::new(&src_lines, *home, &loaded.interns);
let problems = loaded.can_problems.remove(home).unwrap_or_default();
let problems = loaded.can_problems.remove(&home).unwrap_or_default();
for problem in problems.into_iter() {
let report = can_problem(&alloc, module_path.clone(), problem);
let severity = report.severity;
@ -86,25 +71,28 @@ pub fn gen_from_mono_module(
}
}
let problems = loaded.type_problems.remove(&home).unwrap_or_default();
let problems = loaded.type_problems.remove(home).unwrap_or_default();
for problem in problems {
let report = type_problem(&alloc, module_path.clone(), problem);
let severity = report.severity;
let mut buf = String::new();
if let Some(report) = type_problem(&alloc, module_path.clone(), problem) {
let severity = report.severity;
let mut buf = String::new();
report.render_color_terminal(&mut buf, &alloc, &palette);
report.render_color_terminal(&mut buf, &alloc, &palette);
match severity {
Warning => {
warnings.push(buf);
}
RuntimeError => {
errors.push(buf);
match severity {
Warning => {
warnings.push(buf);
}
RuntimeError => {
errors.push(buf);
}
}
}
}
let problems = loaded.mono_problems.remove(&home).unwrap_or_default();
let problems = loaded.mono_problems.remove(home).unwrap_or_default();
for problem in problems {
let report = mono_problem(&alloc, module_path.clone(), problem);
let severity = report.severity;
@ -123,12 +111,18 @@ pub fn gen_from_mono_module(
}
}
let problems_reported;
// Only print warnings if there are no errors
if errors.is_empty() {
problems_reported = warnings.len();
for warning in warnings {
println!("\n{}\n", warning);
}
} else {
problems_reported = errors.len();
for error in errors {
println!("\n{}\n", error);
}
@ -140,10 +134,35 @@ pub fn gen_from_mono_module(
// The horizontal rule is nice when running the program right after
// compiling it, as it lets you clearly see where the compiler
// errors/warnings end and the program output begins.
if total_problems > 0 {
if problems_reported > 0 {
println!("{}\u{001B}[0m\n", Report::horizontal_rule(&palette));
}
problems_reported
}
// TODO how should imported modules factor into this? What if those use builtins too?
// TODO this should probably use more helper functions
// TODO make this polymorphic in the llvm functions so it can be reused for another backend.
#[cfg(feature = "llvm")]
pub fn gen_from_mono_module_llvm(
arena: &bumpalo::Bump,
loaded: MonomorphizedModule,
roc_file_path: &Path,
target: &target_lexicon::Triple,
app_o_file: &Path,
opt_level: OptLevel,
emit_debug_info: bool,
) -> CodeGenTiming {
use crate::target::{self, convert_opt_level};
use inkwell::attributes::{Attribute, AttributeLoc};
use inkwell::context::Context;
use inkwell::module::Linkage;
use inkwell::targets::{CodeModel, FileType, RelocMode};
use std::time::SystemTime;
let code_gen_start = SystemTime::now();
// Generate the binary
let ptr_bytes = target.pointer_width().unwrap().bytes() as u32;
let context = Context::create();
@ -286,6 +305,7 @@ pub fn gen_from_mono_module(
.unwrap();
let llc_args = &[
"-relocation-model=pic",
"-filetype=obj",
app_bc_file.to_str().unwrap(),
"-o",
@ -325,7 +345,7 @@ pub fn gen_from_mono_module(
use target_lexicon::Architecture;
match target.architecture {
Architecture::X86_64 | Architecture::X86_32(_) | Architecture::Aarch64(_) => {
let reloc = RelocMode::Default;
let reloc = RelocMode::PIC;
let model = CodeModel::Default;
let target_machine =
target::target_machine(target, convert_opt_level(opt_level), reloc, model)
@ -354,3 +374,78 @@ pub fn gen_from_mono_module(
emit_o_file,
}
}
pub fn gen_from_mono_module_dev(
arena: &bumpalo::Bump,
loaded: MonomorphizedModule,
target: &target_lexicon::Triple,
app_o_file: &Path,
) -> CodeGenTiming {
use target_lexicon::Architecture;
match target.architecture {
Architecture::Wasm32 => gen_from_mono_module_dev_wasm32(arena, loaded, app_o_file),
Architecture::X86_64 => {
gen_from_mono_module_dev_assembly(arena, loaded, target, app_o_file)
}
_ => todo!(),
}
}
fn gen_from_mono_module_dev_wasm32(
arena: &bumpalo::Bump,
loaded: MonomorphizedModule,
app_o_file: &Path,
) -> CodeGenTiming {
let mut procedures = MutMap::default();
for (key, proc) in loaded.procedures {
procedures.insert(key, proc);
}
let exposed_to_host = loaded
.exposed_to_host
.keys()
.copied()
.collect::<MutSet<_>>();
let env = roc_gen_wasm::Env {
arena,
interns: loaded.interns,
exposed_to_host,
};
let bytes = roc_gen_wasm::build_module(&env, procedures).unwrap();
std::fs::write(&app_o_file, &bytes).expect("failed to write object to file");
CodeGenTiming::default()
}
fn gen_from_mono_module_dev_assembly(
arena: &bumpalo::Bump,
loaded: MonomorphizedModule,
target: &target_lexicon::Triple,
app_o_file: &Path,
) -> CodeGenTiming {
let lazy_literals = true;
let generate_allocators = false; // provided by the platform
let env = roc_gen_dev::Env {
arena,
interns: loaded.interns,
exposed_to_host: loaded.exposed_to_host.keys().copied().collect(),
lazy_literals,
generate_allocators,
};
let module_object = roc_gen_dev::build_module(&env, target, loaded.procedures)
.expect("failed to compile module");
let module_out = module_object
.write()
.expect("failed to build output object");
std::fs::write(&app_o_file, module_out).expect("failed to write object to file");
CodeGenTiming::default()
}

2
compiler/build/src/target.rs
View file

@ -106,7 +106,7 @@ pub fn target_machine(
#[cfg(feature = "llvm")]
pub fn convert_opt_level(level: OptLevel) -> OptimizationLevel {
match level {
OptLevel::Normal => OptimizationLevel::None,
OptLevel::Development | OptLevel::Normal => OptimizationLevel::None,
OptLevel::Optimize => OptimizationLevel::Aggressive,
}
}

8
compiler/builtins/bitcode/src/main.zig
View file

@ -101,6 +101,7 @@ comptime {
exportStrFn(str.strToUtf8C, "to_utf8");
exportStrFn(str.fromUtf8C, "from_utf8");
exportStrFn(str.fromUtf8RangeC, "from_utf8_range");
exportStrFn(str.repeat, "repeat");
}
// Utils
@ -164,7 +165,12 @@ test "" {
// https://github.com/ziglang/zig/blob/85755c51d529e7d9b406c6bdf69ce0a0f33f3353/lib/std/special/compiler_rt/muloti4.zig
//
// Thank you Zig Contributors!
export fn __muloti4(a: i128, b: i128, overflow: *c_int) callconv(.C) i128 {
// Export it as weak incase it is alreadly linked in by something else.
comptime {
@export(__muloti4, .{ .name = "__muloti4", .linkage = .Weak });
}
fn __muloti4(a: i128, b: i128, overflow: *c_int) callconv(.C) i128 {
// @setRuntimeSafety(std.builtin.is_test);
const min = @bitCast(i128, @as(u128, 1 << (128 - 1)));

18
compiler/builtins/bitcode/src/str.zig
View file

@ -866,6 +866,22 @@ pub fn startsWith(string: RocStr, prefix: RocStr) callconv(.C) bool {
return true;
}
// Str.repeat
pub fn repeat(string: RocStr, count: usize) callconv(.C) RocStr {
const bytes_len = string.len();
const bytes_ptr = string.asU8ptr();
var ret_string = RocStr.allocate(.Clone, count * bytes_len);
var ret_string_ptr = ret_string.asU8ptr();
var i: usize = 0;
while (i < count) : (i += 1) {
@memcpy(ret_string_ptr + (i * bytes_len), bytes_ptr, bytes_len);
}
return ret_string;
}
// Str.startsWithCodePt
pub fn startsWithCodePt(string: RocStr, prefix: u32) callconv(.C) bool {
const bytes_ptr = string.asU8ptr();
@ -1150,8 +1166,8 @@ fn strToBytes(arg: RocStr) RocList {
}
const FromUtf8Result = extern struct {
string: RocStr,
byte_index: usize,
string: RocStr,
is_ok: bool,
problem_code: Utf8ByteProblem,
};

1
compiler/builtins/src/bitcode.rs
View file

@ -28,6 +28,7 @@ pub const STR_EQUAL: &str = "roc_builtins.str.equal";
pub const STR_TO_UTF8: &str = "roc_builtins.str.to_utf8";
pub const STR_FROM_UTF8: &str = "roc_builtins.str.from_utf8";
pub const STR_FROM_UTF8_RANGE: &str = "roc_builtins.str.from_utf8_range";
pub const STR_REPEAT: &str = "roc_builtins.str.repeat";
pub const DICT_HASH: &str = "roc_builtins.dict.hash";
pub const DICT_HASH_STR: &str = "roc_builtins.dict.hash_str";

7
compiler/builtins/src/std.rs
View file

@ -618,6 +618,13 @@ pub fn types() -> MutMap<Symbol, (SolvedType, Region)> {
Box::new(str_type())
);
// repeat : Str, Nat -> Str
add_top_level_function_type!(
Symbol::STR_REPEAT,
vec![str_type(), nat_type()],
Box::new(str_type())
);
// fromUtf8 : List U8 -> Result Str [ BadUtf8 Utf8Problem ]*
{
let bad_utf8 = SolvedType::TagUnion(

24
compiler/can/src/annotation.rs
View file

@ -82,6 +82,7 @@ pub fn canonicalize_annotation(
let mut introduced_variables = IntroducedVariables::default();
let mut references = MutSet::default();
let mut aliases = SendMap::default();
let typ = can_annotation_help(
env,
annotation,
@ -249,28 +250,7 @@ fn can_annotation_help(
actual: Box::new(actual),
}
}
None => {
let mut args = Vec::new();
references.insert(symbol);
for arg in *type_arguments {
let arg_ann = can_annotation_help(
env,
&arg.value,
region,
scope,
var_store,
introduced_variables,
local_aliases,
references,
);
args.push(arg_ann);
}
Type::Apply(symbol, args)
}
None => Type::Apply(symbol, args),
}
}
BoundVariable(v) => {

21
compiler/can/src/builtins.rs
View file

@ -66,6 +66,7 @@ pub fn builtin_defs_map(symbol: Symbol, var_store: &mut VarStore) -> Option<Def>
STR_FROM_UTF8_RANGE => str_from_utf8_range,
STR_TO_UTF8 => str_to_utf8,
STR_FROM_FLOAT=> str_from_float,
STR_REPEAT => str_repeat,
LIST_LEN => list_len,
LIST_GET => list_get,
LIST_SET => list_set,
@ -1233,6 +1234,26 @@ fn str_split(symbol: Symbol, var_store: &mut VarStore) -> Def {
)
}
/// Str.repeat : Str, Nat -> Str
fn str_repeat(symbol: Symbol, var_store: &mut VarStore) -> Def {
let str_var = var_store.fresh();
let nat_var = var_store.fresh();
let body = RunLowLevel {
op: LowLevel::StrRepeat,
args: vec![(str_var, Var(Symbol::ARG_1)), (nat_var, Var(Symbol::ARG_2))],
ret_var: str_var,
};
defn(
symbol,
vec![(str_var, Symbol::ARG_1), (nat_var, Symbol::ARG_2)],
var_store,
body,
str_var,
)
}
/// Str.concat : Str, Str -> Str
fn str_concat(symbol: Symbol, var_store: &mut VarStore) -> Def {
let str_var = var_store.fresh();

1
compiler/gen_dev/src/generic64/aarch64.rs
View file

@ -226,6 +226,7 @@ impl CallConv<AArch64GeneralReg, AArch64FloatReg> for AArch64Call {
#[inline(always)]
fn load_args<'a>(
_buf: &mut Vec<'a, u8>,
_symbol_map: &mut MutMap<Symbol, SymbolStorage<AArch64GeneralReg, AArch64FloatReg>>,
_args: &'a [(Layout<'a>, Symbol)],
_ret_layout: &Layout<'a>,

456
compiler/gen_dev/src/generic64/mod.rs
View file

@ -2,15 +2,14 @@ use crate::{Backend, Env, Relocation};
use bumpalo::collections::Vec;
use roc_collections::all::{MutMap, MutSet};
use roc_module::symbol::Symbol;
use roc_mono::ir::{BranchInfo, Literal, Stmt};
use roc_mono::ir::{BranchInfo, JoinPointId, Literal, Param, SelfRecursive, Stmt};
use roc_mono::layout::{Builtin, Layout};
use std::marker::PhantomData;
use target_lexicon::Triple;
pub mod aarch64;
pub mod x86_64;
const PTR_SIZE: u32 = 64;
const PTR_SIZE: u32 = 8;
pub trait CallConv<GeneralReg: RegTrait, FloatReg: RegTrait> {
const GENERAL_PARAM_REGS: &'static [GeneralReg];
@ -49,6 +48,7 @@ pub trait CallConv<GeneralReg: RegTrait, FloatReg: RegTrait> {
// load_args updates the symbol map to know where every arg is stored.
fn load_args<'a>(
buf: &mut Vec<'a, u8>,
symbol_map: &mut MutMap<Symbol, SymbolStorage<GeneralReg, FloatReg>>,
args: &'a [(Layout<'a>, Symbol)],
// ret_layout is needed because if it is a complex type, we pass a pointer as the first arg.
@ -211,12 +211,16 @@ pub struct Backend64Bit<
env: &'a Env<'a>,
buf: Vec<'a, u8>,
relocs: Vec<'a, Relocation>,
proc_name: Option<String>,
is_self_recursive: Option<SelfRecursive>,
last_seen_map: MutMap<Symbol, *const Stmt<'a>>,
layout_map: MutMap<Symbol, *const Layout<'a>>,
layout_map: MutMap<Symbol, Layout<'a>>,
free_map: MutMap<*const Stmt<'a>, Vec<'a, Symbol>>,
symbol_storage_map: MutMap<Symbol, SymbolStorage<GeneralReg, FloatReg>>,
literal_map: MutMap<Symbol, Literal<'a>>,
join_map: MutMap<JoinPointId, u64>,
// This should probably be smarter than a vec.
// There are certain registers we should always use first. With pushing and popping, this could get mixed.
@ -247,11 +251,13 @@ impl<
CC: CallConv<GeneralReg, FloatReg>,
> Backend<'a> for Backend64Bit<'a, GeneralReg, FloatReg, ASM, CC>
{
fn new(env: &'a Env, _target: &Triple) -> Result<Self, String> {
fn new(env: &'a Env) -> Result<Self, String> {
Ok(Backend64Bit {
phantom_asm: PhantomData,
phantom_cc: PhantomData,
env,
proc_name: None,
is_self_recursive: None,
buf: bumpalo::vec![in env.arena],
relocs: bumpalo::vec![in env.arena],
last_seen_map: MutMap::default(),
@ -259,6 +265,7 @@ impl<
free_map: MutMap::default(),
symbol_storage_map: MutMap::default(),
literal_map: MutMap::default(),
join_map: MutMap::default(),
general_free_regs: bumpalo::vec![in env.arena],
general_used_regs: bumpalo::vec![in env.arena],
general_used_callee_saved_regs: MutSet::default(),
@ -275,12 +282,15 @@ impl<
self.env
}
fn reset(&mut self) {
fn reset(&mut self, name: String, is_self_recursive: SelfRecursive) {
self.proc_name = Some(name);
self.is_self_recursive = Some(is_self_recursive);
self.stack_size = 0;
self.free_stack_chunks.clear();
self.fn_call_stack_size = 0;
self.last_seen_map.clear();
self.layout_map.clear();
self.join_map.clear();
self.free_map.clear();
self.symbol_storage_map.clear();
self.buf.clear();
@ -304,7 +314,7 @@ impl<
&mut self.last_seen_map
}
fn layout_map(&mut self) -> &mut MutMap<Symbol, *const Layout<'a>> {
fn layout_map(&mut self) -> &mut MutMap<Symbol, Layout<'a>> {
&mut self.layout_map
}
@ -330,8 +340,49 @@ impl<
)?;
let setup_offset = out.len();
// Deal with jumps to the return address.
let old_relocs = std::mem::replace(&mut self.relocs, bumpalo::vec![in self.env.arena]);
// Check if their is an unnessary jump to return right at the end of the function.
let mut end_jmp_size = 0;
for reloc in old_relocs
.iter()
.filter(|reloc| matches!(reloc, Relocation::JmpToReturn { .. }))
{
if let Relocation::JmpToReturn {
inst_loc,
inst_size,
..
} = reloc
{
if *inst_loc as usize + *inst_size as usize == self.buf.len() {
end_jmp_size = *inst_size as usize;
break;
}
}
}
// Update jumps to returns.
let ret_offset = self.buf.len() - end_jmp_size;
let mut tmp = bumpalo::vec![in self.env.arena];
for reloc in old_relocs
.iter()
.filter(|reloc| matches!(reloc, Relocation::JmpToReturn { .. }))
{
if let Relocation::JmpToReturn {
inst_loc,
inst_size,
offset,
} = reloc
{
if *inst_loc as usize + *inst_size as usize != self.buf.len() {
self.update_jmp_imm32_offset(&mut tmp, *inst_loc, *offset, ret_offset as u64);
}
}
}
// Add function body.
out.extend(&self.buf);
out.extend(&self.buf[..self.buf.len() - end_jmp_size]);
// Cleanup stack.
CC::cleanup_stack(
@ -342,23 +393,28 @@ impl<
)?;
ASM::ret(&mut out);
// Update relocs to include stack setup offset.
// Update other relocs to include stack setup offset.
let mut out_relocs = bumpalo::vec![in self.env.arena];
let old_relocs = std::mem::replace(&mut self.relocs, bumpalo::vec![in self.env.arena]);
out_relocs.extend(old_relocs.into_iter().map(|reloc| match reloc {
Relocation::LocalData { offset, data } => Relocation::LocalData {
offset: offset + setup_offset as u64,
data,
},
Relocation::LinkedData { offset, name } => Relocation::LinkedData {
offset: offset + setup_offset as u64,
name,
},
Relocation::LinkedFunction { offset, name } => Relocation::LinkedFunction {
offset: offset + setup_offset as u64,
name,
},
}));
out_relocs.extend(
old_relocs
.into_iter()
.filter(|reloc| !matches!(reloc, Relocation::JmpToReturn { .. }))
.map(|reloc| match reloc {
Relocation::LocalData { offset, data } => Relocation::LocalData {
offset: offset + setup_offset as u64,
data,
},
Relocation::LinkedData { offset, name } => Relocation::LinkedData {
offset: offset + setup_offset as u64,
name,
},
Relocation::LinkedFunction { offset, name } => Relocation::LinkedFunction {
offset: offset + setup_offset as u64,
name,
},
Relocation::JmpToReturn { .. } => unreachable!(),
}),
);
Ok((out.into_bump_slice(), out_relocs.into_bump_slice()))
}
@ -367,7 +423,12 @@ impl<
args: &'a [(Layout<'a>, Symbol)],
ret_layout: &Layout<'a>,
) -> Result<(), String> {
CC::load_args(&mut self.symbol_storage_map, args, ret_layout)?;
CC::load_args(
&mut self.buf,
&mut self.symbol_storage_map,
args,
ret_layout,
)?;
// Update used and free regs.
for (sym, storage) in &self.symbol_storage_map {
match storage {
@ -401,29 +462,13 @@ impl<
arg_layouts: &[Layout<'a>],
ret_layout: &Layout<'a>,
) -> Result<(), String> {
if let Some(SelfRecursive::SelfRecursive(id)) = self.is_self_recursive {
if &fn_name == self.proc_name.as_ref().unwrap() && self.join_map.contains_key(&id) {
return self.build_jump(&id, args, arg_layouts, ret_layout);
}
}
// Save used caller saved regs.
let old_general_used_regs = std::mem::replace(
&mut self.general_used_regs,
bumpalo::vec![in self.env.arena],
);
for (reg, saved_sym) in old_general_used_regs.into_iter() {
if CC::general_caller_saved(&reg) {
self.general_free_regs.push(reg);
self.free_to_stack(&saved_sym)?;
} else {
self.general_used_regs.push((reg, saved_sym));
}
}
let old_float_used_regs =
std::mem::replace(&mut self.float_used_regs, bumpalo::vec![in self.env.arena]);
for (reg, saved_sym) in old_float_used_regs.into_iter() {
if CC::float_caller_saved(&reg) {
self.float_free_regs.push(reg);
self.free_to_stack(&saved_sym)?;
} else {
self.float_used_regs.push((reg, saved_sym));
}
}
self.push_used_caller_saved_regs_to_stack()?;
// Put values in param regs or on top of the stack.
let tmp_stack_size = CC::store_args(
@ -450,6 +495,25 @@ impl<
ASM::mov_freg64_freg64(&mut self.buf, dst_reg, CC::FLOAT_RETURN_REGS[0]);
Ok(())
}
Layout::Builtin(Builtin::Str) => {
if CC::returns_via_arg_pointer(ret_layout)? {
// This will happen on windows, return via pointer here.
Err("FnCall: Returning strings via pointer not yet implemented".to_string())
} else {
let offset = self.claim_stack_size(16)?;
self.symbol_storage_map.insert(
*dst,
SymbolStorage::Base {
offset,
size: 16,
owned: true,
},
);
ASM::mov_base32_reg64(&mut self.buf, offset, CC::GENERAL_RETURN_REGS[0]);
ASM::mov_base32_reg64(&mut self.buf, offset + 8, CC::GENERAL_RETURN_REGS[1]);
Ok(())
}
}
x => Err(format!(
"FnCall: receiving return type, {:?}, is not yet implemented",
x
@ -486,7 +550,7 @@ impl<
// Build unconditional jump to the end of this switch.
// Since we don't know the offset yet, set it to 0 and overwrite later.
let jmp_location = self.buf.len();
let jmp_offset = ASM::jmp_imm32(&mut self.buf, 0);
let jmp_offset = ASM::jmp_imm32(&mut self.buf, 0x1234_5678);
ret_jumps.push((jmp_location, jmp_offset));
// Overwite the original jne with the correct offset.
@ -510,12 +574,12 @@ impl<
// Update all return jumps to jump past the default case.
let ret_offset = self.buf.len();
for (jmp_location, start_offset) in ret_jumps.into_iter() {
tmp.clear();
let jmp_offset = ret_offset - start_offset;
ASM::jmp_imm32(&mut tmp, jmp_offset as i32);
for (i, byte) in tmp.iter().enumerate() {
self.buf[jmp_location + i] = *byte;
}
self.update_jmp_imm32_offset(
&mut tmp,
jmp_location as u64,
start_offset as u64,
ret_offset as u64,
);
}
Ok(())
} else {
@ -526,6 +590,134 @@ impl<
}
}
fn build_join(
&mut self,
id: &JoinPointId,
parameters: &'a [Param<'a>],
body: &'a Stmt<'a>,
remainder: &'a Stmt<'a>,
ret_layout: &Layout<'a>,
) -> Result<(), String> {
// Create jump to remaining.
let jmp_location = self.buf.len();
let start_offset = ASM::jmp_imm32(&mut self.buf, 0x1234_5678);
// This section can essentially be seen as a sub function within the main function.
// Thus we build using a new backend with some minor extra synchronization.
let mut sub_backend = Self::new(self.env)?;
sub_backend.reset(
self.proc_name.as_ref().unwrap().clone(),
self.is_self_recursive.as_ref().unwrap().clone(),
);
// Sync static maps of important information.
sub_backend.last_seen_map = self.last_seen_map.clone();
sub_backend.layout_map = self.layout_map.clone();
sub_backend.free_map = self.free_map.clone();
// Setup join point.
sub_backend.join_map.insert(*id, 0);
self.join_map.insert(*id, self.buf.len() as u64);
// Sync stack size so the "sub function" doesn't mess up our stack.
sub_backend.stack_size = self.stack_size;
sub_backend.fn_call_stack_size = self.fn_call_stack_size;
// Load params as if they were args.
let mut args = bumpalo::vec![in self.env.arena];
for param in parameters {
args.push((param.layout, param.symbol));
}
sub_backend.load_args(args.into_bump_slice(), ret_layout)?;
// Build all statements in body.
sub_backend.build_stmt(body, ret_layout)?;
// Merge the "sub function" into the main function.
let sub_func_offset = self.buf.len() as u64;
self.buf.extend_from_slice(&sub_backend.buf);
// Update stack based on how much was used by the sub function.
self.stack_size = sub_backend.stack_size;
self.fn_call_stack_size = sub_backend.fn_call_stack_size;
// Relocations must be shifted to be merged correctly.
self.relocs
.extend(sub_backend.relocs.into_iter().map(|reloc| match reloc {
Relocation::LocalData { offset, data } => Relocation::LocalData {
offset: offset + sub_func_offset,
data,
},
Relocation::LinkedData { offset, name } => Relocation::LinkedData {
offset: offset + sub_func_offset,
name,
},
Relocation::LinkedFunction { offset, name } => Relocation::LinkedFunction {
offset: offset + sub_func_offset,
name,
},
Relocation::JmpToReturn {
inst_loc,
inst_size,
offset,
} => Relocation::JmpToReturn {
inst_loc: inst_loc + sub_func_offset,
inst_size,
offset: offset + sub_func_offset,
},
}));
// Overwite the original jump with the correct offset.
let mut tmp = bumpalo::vec![in self.env.arena];
self.update_jmp_imm32_offset(
&mut tmp,
jmp_location as u64,
start_offset as u64,
self.buf.len() as u64,
);
// Build remainder of function.
self.build_stmt(remainder, ret_layout)
}
fn build_jump(
&mut self,
id: &JoinPointId,
args: &'a [Symbol],
arg_layouts: &[Layout<'a>],
ret_layout: &Layout<'a>,
) -> Result<(), String> {
// Treat this like a function call, but with a jump instead of a call instruction at the end.
self.push_used_caller_saved_regs_to_stack()?;
let tmp_stack_size = CC::store_args(
&mut self.buf,
&self.symbol_storage_map,
args,
arg_layouts,
ret_layout,
)?;
self.fn_call_stack_size = std::cmp::max(self.fn_call_stack_size, tmp_stack_size);
let jmp_location = self.buf.len();
let start_offset = ASM::jmp_imm32(&mut self.buf, 0x1234_5678);
if let Some(offset) = self.join_map.get(id) {
let offset = *offset;
let mut tmp = bumpalo::vec![in self.env.arena];
self.update_jmp_imm32_offset(
&mut tmp,
jmp_location as u64,
start_offset as u64,
offset,
);
Ok(())
} else {
Err(format!(
"Jump: unknown point specified to jump to: {:?}",
id
))
}
}
fn build_num_abs(
&mut self,
dst: &Symbol,
@ -726,6 +918,35 @@ impl<
ASM::mov_freg64_imm64(&mut self.buf, &mut self.relocs, reg, val);
Ok(())
}
Literal::Str(x) if x.len() < 16 => {
// Load small string.
let reg = self.get_tmp_general_reg()?;
let offset = self.claim_stack_size(16)?;
self.symbol_storage_map.insert(
*sym,
SymbolStorage::Base {
offset,
size: 16,
owned: true,
},
);
let mut bytes = [0; 16];
bytes[..x.len()].copy_from_slice(x.as_bytes());
bytes[15] = (x.len() as u8) | 0b1000_0000;
let mut num_bytes = [0; 8];
num_bytes.copy_from_slice(&bytes[..8]);
let num = i64::from_ne_bytes(num_bytes);
ASM::mov_reg64_imm64(&mut self.buf, reg, num);
ASM::mov_base32_reg64(&mut self.buf, offset, reg);
num_bytes.copy_from_slice(&bytes[8..]);
let num = i64::from_ne_bytes(num_bytes);
ASM::mov_reg64_imm64(&mut self.buf, reg, num);
ASM::mov_base32_reg64(&mut self.buf, offset + 8, reg);
Ok(())
}
x => Err(format!("loading literal, {:?}, is not yet implemented", x)),
}
}
@ -828,29 +1049,39 @@ impl<
fn return_symbol(&mut self, sym: &Symbol, layout: &Layout<'a>) -> Result<(), String> {
let val = self.symbol_storage_map.get(sym);
match val {
Some(SymbolStorage::GeneralReg(reg)) if *reg == CC::GENERAL_RETURN_REGS[0] => Ok(()),
Some(SymbolStorage::GeneralReg(reg)) if *reg == CC::GENERAL_RETURN_REGS[0] => {}
Some(SymbolStorage::GeneralReg(reg)) => {
// If it fits in a general purpose register, just copy it over to.
// Technically this can be optimized to produce shorter instructions if less than 64bits.
ASM::mov_reg64_reg64(&mut self.buf, CC::GENERAL_RETURN_REGS[0], *reg);
Ok(())
}
Some(SymbolStorage::FloatReg(reg)) if *reg == CC::FLOAT_RETURN_REGS[0] => Ok(()),
Some(SymbolStorage::FloatReg(reg)) if *reg == CC::FLOAT_RETURN_REGS[0] => {}
Some(SymbolStorage::FloatReg(reg)) => {
ASM::mov_freg64_freg64(&mut self.buf, CC::FLOAT_RETURN_REGS[0], *reg);
Ok(())
}
Some(SymbolStorage::Base { offset, size, .. }) => match layout {
Layout::Builtin(Builtin::Int64) => {
ASM::mov_reg64_base32(&mut self.buf, CC::GENERAL_RETURN_REGS[0], *offset);
Ok(())
}
Layout::Builtin(Builtin::Float64) => {
ASM::mov_freg64_base32(&mut self.buf, CC::FLOAT_RETURN_REGS[0], *offset);
Ok(())
}
Layout::Builtin(Builtin::Str) => {
if self.symbol_storage_map.contains_key(&Symbol::RET_POINTER) {
// This will happen on windows, return via pointer here.
return Err("Returning strings via pointer not yet implemented".to_string());
} else {
ASM::mov_reg64_base32(&mut self.buf, CC::GENERAL_RETURN_REGS[0], *offset);
ASM::mov_reg64_base32(
&mut self.buf,
CC::GENERAL_RETURN_REGS[1],
*offset + 8,
);
}
}
Layout::Struct(field_layouts) => {
let (offset, size) = (*offset, *size);
// Nothing to do for empty struct
if size > 0 {
let ret_reg = if self.symbol_storage_map.contains_key(&Symbol::RET_POINTER)
{
@ -858,23 +1089,34 @@ impl<
} else {
None
};
CC::return_struct(&mut self.buf, offset, size, field_layouts, ret_reg)
} else {
// Nothing to do for empty struct
Ok(())
CC::return_struct(&mut self.buf, offset, size, field_layouts, ret_reg)?;
}
}
x => Err(format!(
"returning symbol with layout, {:?}, is not yet implemented",
x
)),
x => {
return Err(format!(
"returning symbol with layout, {:?}, is not yet implemented",
x
));
}
},
Some(x) => Err(format!(
"returning symbol storage, {:?}, is not yet implemented",
x
)),
None => Err(format!("Unknown return symbol: {}", sym)),
Some(x) => {
return Err(format!(
"returning symbol storage, {:?}, is not yet implemented",
x
));
}
None => {
return Err(format!("Unknown return symbol: {}", sym));
}
}
let inst_loc = self.buf.len() as u64;
let offset = ASM::jmp_imm32(&mut self.buf, 0x1234_5678) as u64;
self.relocs.push(Relocation::JmpToReturn {
inst_loc,
inst_size: self.buf.len() as u64 - inst_loc,
offset,
});
Ok(())
}
}
@ -1212,4 +1454,72 @@ impl<
)),
}
}
fn push_used_caller_saved_regs_to_stack(&mut self) -> Result<(), String> {
let old_general_used_regs = std::mem::replace(
&mut self.general_used_regs,
bumpalo::vec![in self.env.arena],
);
for (reg, saved_sym) in old_general_used_regs.into_iter() {
if CC::general_caller_saved(&reg) {
self.general_free_regs.push(reg);
self.free_to_stack(&saved_sym)?;
} else {
self.general_used_regs.push((reg, saved_sym));
}
}
let old_float_used_regs =
std::mem::replace(&mut self.float_used_regs, bumpalo::vec![in self.env.arena]);
for (reg, saved_sym) in old_float_used_regs.into_iter() {
if CC::float_caller_saved(&reg) {
self.float_free_regs.push(reg);
self.free_to_stack(&saved_sym)?;
} else {
self.float_used_regs.push((reg, saved_sym));
}
}
Ok(())
}
// Updates a jump instruction to a new offset and returns the number of bytes written.
fn update_jmp_imm32_offset(
&mut self,
tmp: &mut Vec<'a, u8>,
jmp_location: u64,
base_offset: u64,
target_offset: u64,
) {
tmp.clear();
let jmp_offset = target_offset as i32 - base_offset as i32;
ASM::jmp_imm32(tmp, jmp_offset);
for (i, byte) in tmp.iter().enumerate() {
self.buf[jmp_location as usize + i] = *byte;
}
}
}
#[macro_export]
macro_rules! single_register_integers {
() => {
Builtin::Int1
| Builtin::Int8
| Builtin::Int16
| Builtin::Int32
| Builtin::Int64
| Builtin::Usize
};
}
#[macro_export]
macro_rules! single_register_floats {
() => {
Builtin::Float32 | Builtin::Float64
};
}
#[macro_export]
macro_rules! single_register_builtins {
() => {
single_register_integers!() | single_register_floats!()
};
}

107
compiler/gen_dev/src/generic64/x86_64.rs
View file

@ -1,5 +1,7 @@
use crate::generic64::{Assembler, CallConv, RegTrait, SymbolStorage, PTR_SIZE};
use crate::Relocation;
use crate::{
single_register_builtins, single_register_floats, single_register_integers, Relocation,
};
use bumpalo::collections::Vec;
use roc_collections::all::MutMap;
use roc_module::symbol::Symbol;
@ -175,6 +177,7 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64SystemV {
#[inline(always)]
fn load_args<'a>(
buf: &mut Vec<'a, u8>,
symbol_map: &mut MutMap<Symbol, SymbolStorage<X86_64GeneralReg, X86_64FloatReg>>,
args: &'a [(Layout<'a>, Symbol)],
ret_layout: &Layout<'a>,
@ -191,7 +194,7 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64SystemV {
}
for (layout, sym) in args.iter() {
match layout {
Layout::Builtin(Builtin::Int64) => {
Layout::Builtin(single_register_integers!()) => {
if general_i < Self::GENERAL_PARAM_REGS.len() {
symbol_map.insert(
*sym,
@ -210,7 +213,7 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64SystemV {
);
}
}
Layout::Builtin(Builtin::Float64) => {
Layout::Builtin(single_register_floats!()) => {
if float_i < Self::FLOAT_PARAM_REGS.len() {
symbol_map.insert(
*sym,
@ -229,6 +232,30 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64SystemV {
);
}
}
Layout::Builtin(Builtin::Str) => {
if general_i + 1 < Self::GENERAL_PARAM_REGS.len() {
// Load the value to the param reg.
let dst1 = Self::GENERAL_PARAM_REGS[general_i];
let dst2 = Self::GENERAL_PARAM_REGS[general_i + 1];
base_offset += 16;
X86_64Assembler::mov_reg64_base32(buf, dst1, base_offset - 8);
X86_64Assembler::mov_reg64_base32(buf, dst2, base_offset);
symbol_map.insert(
*sym,
SymbolStorage::Base {
offset: base_offset,
size: 16,
owned: true,
},
);
general_i += 2;
} else {
return Err(
"loading strings args on the stack is not yet implemented".to_string()
);
}
}
Layout::Struct(&[]) => {}
x => {
return Err(format!(
"Loading args with layout {:?} not yet implementd",
@ -254,8 +281,7 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64SystemV {
// For most return layouts we will do nothing.
// In some cases, we need to put the return address as the first arg.
match ret_layout {
Layout::Builtin(Builtin::Int64) => {}
Layout::Builtin(Builtin::Float64) => {}
Layout::Builtin(single_register_builtins!() | Builtin::Str) => {}
x => {
return Err(format!(
"receiving return type, {:?}, is not yet implemented",
@ -265,7 +291,7 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64SystemV {
}
for (i, layout) in arg_layouts.iter().enumerate() {
match layout {
Layout::Builtin(Builtin::Int64) => {
Layout::Builtin(single_register_integers!()) => {
if general_i < Self::GENERAL_PARAM_REGS.len() {
// Load the value to the param reg.
let dst = Self::GENERAL_PARAM_REGS[general_i];
@ -319,7 +345,7 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64SystemV {
stack_offset += 8;
}
}
Layout::Builtin(Builtin::Float64) => {
Layout::Builtin(single_register_floats!()) => {
if float_i < Self::FLOAT_PARAM_REGS.len() {
// Load the value to the param reg.
let dst = Self::FLOAT_PARAM_REGS[float_i];
@ -371,6 +397,33 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64SystemV {
stack_offset += 8;
}
}
Layout::Builtin(Builtin::Str) => {
if general_i + 1 < Self::GENERAL_PARAM_REGS.len() {
// Load the value to the param reg.
let dst1 = Self::GENERAL_PARAM_REGS[general_i];
let dst2 = Self::GENERAL_PARAM_REGS[general_i + 1];
match symbol_map
.get(&args[i])
.ok_or("function argument does not reference any symbol")?
{
SymbolStorage::Base { offset, .. } => {
X86_64Assembler::mov_reg64_base32(buf, dst1, *offset);
X86_64Assembler::mov_reg64_base32(buf, dst2, *offset + 8);
}
_ => {
return Err("Strings only support being loaded from base offsets"
.to_string());
}
}
general_i += 2;
} else {
return Err(
"calling functions with strings on the stack is not yet implemented"
.to_string(),
);
}
}
Layout::Struct(&[]) => {}
x => {
return Err(format!(
"calling with arg type, {:?}, is not yet implemented",
@ -513,6 +566,7 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64WindowsFastcall {
#[inline(always)]
fn load_args<'a>(
_buf: &mut Vec<'a, u8>,
symbol_map: &mut MutMap<Symbol, SymbolStorage<X86_64GeneralReg, X86_64FloatReg>>,
args: &'a [(Layout<'a>, Symbol)],
ret_layout: &Layout<'a>,
@ -529,13 +583,23 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64WindowsFastcall {
for (layout, sym) in args.iter() {
if i < Self::GENERAL_PARAM_REGS.len() {
match layout {
Layout::Builtin(Builtin::Int64) => {
Layout::Builtin(single_register_integers!()) => {
symbol_map
.insert(*sym, SymbolStorage::GeneralReg(Self::GENERAL_PARAM_REGS[i]));
i += 1;
}
Layout::Builtin(Builtin::Float64) => {
Layout::Builtin(single_register_floats!()) => {
symbol_map.insert(*sym, SymbolStorage::FloatReg(Self::FLOAT_PARAM_REGS[i]));
i += 1;
}
Layout::Builtin(Builtin::Str) => {
// I think this just needs to be passed on the stack, so not a huge deal.
return Err(
"Passing str args with Windows fast call not yet implemented."
.to_string(),
);
}
Layout::Struct(&[]) => {}
x => {
return Err(format!(
"Loading args with layout {:?} not yet implementd",
@ -543,11 +607,9 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64WindowsFastcall {
));
}
}
i += 1;
} else {
base_offset += match layout {
Layout::Builtin(Builtin::Int64) => 8,
Layout::Builtin(Builtin::Float64) => 8,
Layout::Builtin(single_register_builtins!()) => 8,
x => {
return Err(format!(
"Loading args with layout {:?} not yet implemented",
@ -577,12 +639,10 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64WindowsFastcall {
ret_layout: &Layout<'a>,
) -> Result<u32, String> {
let mut stack_offset = Self::SHADOW_SPACE_SIZE as i32;
let mut reg_i = 0;
// For most return layouts we will do nothing.
// In some cases, we need to put the return address as the first arg.
match ret_layout {
Layout::Builtin(Builtin::Int64) => {}
Layout::Builtin(Builtin::Float64) => {}
Layout::Builtin(single_register_builtins!()) => {}
x => {
return Err(format!(
"receiving return type, {:?}, is not yet implemented",
@ -592,10 +652,10 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64WindowsFastcall {
}
for (i, layout) in arg_layouts.iter().enumerate() {
match layout {
Layout::Builtin(Builtin::Int64) => {
Layout::Builtin(single_register_integers!()) => {
if i < Self::GENERAL_PARAM_REGS.len() {
// Load the value to the param reg.
let dst = Self::GENERAL_PARAM_REGS[reg_i];
let dst = Self::GENERAL_PARAM_REGS[i];
match symbol_map
.get(&args[i])
.ok_or("function argument does not reference any symbol")?
@ -613,7 +673,6 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64WindowsFastcall {
)
}
}
reg_i += 1;
} else {
// Load the value to the stack.
match symbol_map
@ -646,10 +705,10 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64WindowsFastcall {
stack_offset += 8;
}
}
Layout::Builtin(Builtin::Float64) => {
Layout::Builtin(single_register_floats!()) => {
if i < Self::FLOAT_PARAM_REGS.len() {
// Load the value to the param reg.
let dst = Self::FLOAT_PARAM_REGS[reg_i];
let dst = Self::FLOAT_PARAM_REGS[i];
match symbol_map
.get(&args[i])
.ok_or("function argument does not reference any symbol")?
@ -666,7 +725,6 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64WindowsFastcall {
return Err("Cannot load general symbol into FloatReg".to_string())
}
}
reg_i += 1;
} else {
// Load the value to the stack.
match symbol_map
@ -698,6 +756,13 @@ impl CallConv<X86_64GeneralReg, X86_64FloatReg> for X86_64WindowsFastcall {
stack_offset += 8;
}
}
Layout::Builtin(Builtin::Str) => {
// I think this just needs to be passed on the stack, so not a huge deal.
return Err(
"Passing str args with Windows fast call not yet implemented.".to_string(),
);
}
Layout::Struct(&[]) => {}
x => {
return Err(format!(
"calling with arg type, {:?}, is not yet implemented",

103
compiler/gen_dev/src/lib.rs
View file

@ -9,10 +9,10 @@ use roc_module::ident::{ModuleName, TagName};
use roc_module::low_level::LowLevel;
use roc_module::symbol::{Interns, Symbol};
use roc_mono::ir::{
BranchInfo, CallType, Expr, JoinPointId, ListLiteralElement, Literal, Proc, Stmt,
BranchInfo, CallType, Expr, JoinPointId, ListLiteralElement, Literal, Param, Proc,
SelfRecursive, Stmt,
};
use roc_mono::layout::{Builtin, Layout, LayoutIds};
use target_lexicon::Triple;
mod generic64;
mod object_builder;
@ -46,6 +46,11 @@ pub enum Relocation {
offset: u64,
name: String,
},
JmpToReturn {
inst_loc: u64,
inst_size: u64,
offset: u64,
},
}
trait Backend<'a>
@ -53,12 +58,13 @@ where
Self: Sized,
{
/// new creates a new backend that will output to the specific Object.
fn new(env: &'a Env, target: &Triple) -> Result<Self, String>;
fn new(env: &'a Env) -> Result<Self, String>;
fn env(&self) -> &'a Env<'a>;
/// reset resets any registers or other values that may be occupied at the end of a procedure.
fn reset(&mut self);
/// It also passes basic procedure information to the builder for setup of the next function.
fn reset(&mut self, name: String, is_self_recursive: SelfRecursive);
/// finalize does any setup and cleanup that should happen around the procedure.
/// finalize does setup because things like stack size and jump locations are not know until the function is written.
@ -79,17 +85,16 @@ where
/// build_proc creates a procedure and outputs it to the wrapped object writer.
fn build_proc(&mut self, proc: Proc<'a>) -> Result<(&'a [u8], &[Relocation]), String> {
self.reset();
let proc_name = LayoutIds::default()
.get(proc.name, &proc.ret_layout)
.to_symbol_string(proc.name, &self.env().interns);
self.reset(proc_name, proc.is_self_recursive);
self.load_args(proc.args, &proc.ret_layout)?;
for (layout, sym) in proc.args {
self.set_layout_map(*sym, layout)?;
}
// let start = std::time::Instant::now();
self.scan_ast(&proc.body);
self.create_free_map();
// let duration = start.elapsed();
// println!("Time to calculate lifetimes: {:?}", duration);
// println!("{:?}", self.last_seen_map());
self.build_stmt(&proc.body, &proc.ret_layout)?;
self.finalize()
}
@ -110,6 +115,11 @@ where
self.free_symbols(stmt)?;
Ok(())
}
Stmt::Refcounting(_modify, following) => {
// TODO: actually deal with refcounting. For hello world, we just skipped it.
self.build_stmt(following, ret_layout)?;
Ok(())
}
Stmt::Switch {
cond_symbol,
cond_layout,
@ -128,6 +138,35 @@ where
self.free_symbols(stmt)?;
Ok(())
}
Stmt::Join {
id,
parameters,
body,
remainder,
} => {
for param in parameters.iter() {
self.set_layout_map(param.symbol, &param.layout)?;
}
self.build_join(id, parameters, body, remainder, ret_layout)?;
self.free_symbols(stmt)?;
Ok(())
}
Stmt::Jump(id, args) => {
let mut arg_layouts: bumpalo::collections::Vec<Layout<'a>> =
bumpalo::vec![in self.env().arena];
arg_layouts.reserve(args.len());
let layout_map = self.layout_map();
for arg in *args {
if let Some(layout) = layout_map.get(arg) {
arg_layouts.push(*layout);
} else {
return Err(format!("the argument, {:?}, has no know layout", arg));
}
}
self.build_jump(id, args, arg_layouts.into_bump_slice(), ret_layout)?;
self.free_symbols(stmt)?;
Ok(())
}
x => Err(format!("the statement, {:?}, is not yet implemented", x)),
}
}
@ -141,6 +180,25 @@ where
ret_layout: &Layout<'a>,
) -> Result<(), String>;
// build_join generates a instructions for a join statement.
fn build_join(
&mut self,
id: &JoinPointId,
parameters: &'a [Param<'a>],
body: &'a Stmt<'a>,
remainder: &'a Stmt<'a>,
ret_layout: &Layout<'a>,
) -> Result<(), String>;
// build_jump generates a instructions for a jump statement.
fn build_jump(
&mut self,
id: &JoinPointId,
args: &'a [Symbol],
arg_layouts: &[Layout<'a>],
ret_layout: &Layout<'a>,
) -> Result<(), String>;
/// build_expr builds the expressions for the specified symbol.
/// The builder must keep track of the symbol because it may be referred to later.
fn build_expr(
@ -241,6 +299,13 @@ where
arg_layouts,
ret_layout,
),
Symbol::STR_CONCAT => self.build_run_low_level(
sym,
&LowLevel::StrConcat,
arguments,
arg_layouts,
ret_layout,
),
x if x
.module_string(&self.env().interns)
.starts_with(ModuleName::APP) =>
@ -263,8 +328,7 @@ where
let layout_map = self.layout_map();
for arg in *arguments {
if let Some(layout) = layout_map.get(arg) {
// This is safe because every value in the map is always set with a valid layout and cannot be null.
arg_layouts.push(unsafe { *(*layout) });
arg_layouts.push(*layout);
} else {
return Err(format!("the argument, {:?}, has no know layout", arg));
}
@ -414,6 +478,13 @@ where
arg_layouts,
ret_layout,
),
LowLevel::StrConcat => self.build_fn_call(
sym,
bitcode::STR_CONCAT.to_string(),
args,
arg_layouts,
ret_layout,
),
x => Err(format!("low level, {:?}. is not yet implemented", x)),
}
}
@ -507,7 +578,7 @@ where
/// load_literal sets a symbol to be equal to a literal.
fn load_literal(&mut self, sym: &Symbol, lit: &Literal<'a>) -> Result<(), String>;
/// return_symbol moves a symbol to the correct return location for the backend.
/// return_symbol moves a symbol to the correct return location for the backend and adds a jump to the end of the function.
fn return_symbol(&mut self, sym: &Symbol, layout: &Layout<'a>) -> Result<(), String>;
/// free_symbols will free all symbols for the given statement.
@ -542,12 +613,10 @@ where
/// set_layout_map sets the layout for a specific symbol.
fn set_layout_map(&mut self, sym: Symbol, layout: &Layout<'a>) -> Result<(), String> {
if let Some(x) = self.layout_map().insert(sym, layout) {
if let Some(old_layout) = self.layout_map().insert(sym, *layout) {
// Layout map already contains the symbol. We should never need to overwrite.
// If the layout is not the same, that is a bug.
// There is always an old layout value and this dereference is safe.
let old_layout = unsafe { *x };
if old_layout != *layout {
if &old_layout != layout {
Err(format!(
"Overwriting layout for symbol, {:?}. This should never happen. got {:?}, want {:?}",
sym, layout, old_layout
@ -561,7 +630,7 @@ where
}
/// layout_map gets the map from symbol to layout.
fn layout_map(&mut self) -> &mut MutMap<Symbol, *const Layout<'a>>;
fn layout_map(&mut self) -> &mut MutMap<Symbol, Layout<'a>>;
fn create_free_map(&mut self) {
let mut free_map = MutMap::default();

15
compiler/gen_dev/src/object_builder.rs
View file

@ -34,7 +34,7 @@ pub fn build_module<'a>(
x86_64::X86_64FloatReg,
x86_64::X86_64Assembler,
x86_64::X86_64SystemV,
> = Backend::new(env, target)?;
> = Backend::new(env)?;
build_object(
env,
procedures,
@ -52,7 +52,7 @@ pub fn build_module<'a>(
x86_64::X86_64FloatReg,
x86_64::X86_64Assembler,
x86_64::X86_64SystemV,
> = Backend::new(env, target)?;
> = Backend::new(env)?;
build_object(
env,
procedures,
@ -74,7 +74,7 @@ pub fn build_module<'a>(
aarch64::AArch64FloatReg,
aarch64::AArch64Assembler,
aarch64::AArch64Call,
> = Backend::new(env, target)?;
> = Backend::new(env)?;
build_object(
env,
procedures,
@ -191,10 +191,16 @@ fn build_object<'a, B: Backend<'a>>(
let mut layout_ids = roc_mono::layout::LayoutIds::default();
let mut procs = Vec::with_capacity_in(procedures.len(), env.arena);
for ((sym, layout), proc) in procedures {
let fn_name = layout_ids
let base_name = layout_ids
.get_toplevel(sym, &layout)
.to_symbol_string(sym, &env.interns);
let fn_name = if env.exposed_to_host.contains(&sym) {
format!("roc_{}_exposed", base_name)
} else {
base_name
};
let section_id = output.add_section(
output.segment_name(StandardSegment::Text).to_vec(),
format!(".text.{:x}", sym.as_u64()).as_bytes().to_vec(),
@ -298,6 +304,7 @@ fn build_object<'a, B: Backend<'a>>(
return Err(format!("failed to find fn symbol for {:?}", name));
}
}
Relocation::JmpToReturn { .. } => unreachable!(),
};
relocations.push((section_id, elfreloc));
}

144
compiler/gen_dev/tests/dev_num.rs
View file

@ -1,12 +1,6 @@
#[macro_use]
extern crate pretty_assertions;
#[macro_use]
extern crate indoc;
extern crate bumpalo;
extern crate libc;
#[macro_use]
mod helpers;
@ -281,6 +275,24 @@ mod dev_num {
);
}
#[test]
fn gen_fast_fib_fn() {
assert_evals_to!(
indoc!(
r#"
fib = \n, a, b ->
if n == 0 then
a
else
fib (n - 1) b (a + b)
fib 10 0 1
"#
),
55,
i64
);
}
#[test]
fn f64_abs() {
assert_evals_to!("Num.abs -4.7", 4.7, f64);
@ -580,18 +592,18 @@ mod dev_num {
// assert_evals_to!("0.0 >= 0.0", true, bool);
// }
// #[test]
// fn gen_order_of_arithmetic_ops() {
// assert_evals_to!(
// indoc!(
// r#"
// 1 + 3 * 7 - 2
// "#
// ),
// 20,
// i64
// );
// }
#[test]
fn gen_order_of_arithmetic_ops() {
assert_evals_to!(
indoc!(
r#"
1 + 3 * 7 - 2
"#
),
20,
i64
);
}
// #[test]
// fn gen_order_of_arithmetic_ops_complex_float() {
@ -606,59 +618,59 @@ mod dev_num {
// );
// }
// #[test]
// fn if_guard_bind_variable_false() {
// assert_evals_to!(
// indoc!(
// r#"
// wrapper = \{} ->
// when 10 is
// x if x == 5 -> 0
// _ -> 42
#[test]
fn if_guard_bind_variable_false() {
assert_evals_to!(
indoc!(
r#"
wrapper = \{} ->
when 10 is
x if x == 5 -> 0
_ -> 42
// wrapper {}
// "#
// ),
// 42,
// i64
// );
// }
wrapper {}
"#
),
42,
i64
);
}
// #[test]
// fn if_guard_bind_variable_true() {
// assert_evals_to!(
// indoc!(
// r#"
// wrapper = \{} ->
// when 10 is
// x if x == 10 -> 42
// _ -> 0
#[test]
fn if_guard_bind_variable_true() {
assert_evals_to!(
indoc!(
r#"
wrapper = \{} ->
when 10 is
x if x == 10 -> 42
_ -> 0
// wrapper {}
// "#
// ),
// 42,
// i64
// );
// }
wrapper {}
"#
),
42,
i64
);
}
// #[test]
// fn tail_call_elimination() {
// assert_evals_to!(
// indoc!(
// r#"
// sum = \n, accum ->
// when n is
// 0 -> accum
// _ -> sum (n - 1) (n + accum)
#[test]
fn tail_call_elimination() {
assert_evals_to!(
indoc!(
r#"
sum = \n, accum ->
when n is
0 -> accum
_ -> sum (n - 1) (n + accum)
// sum 1_000_000 0
// "#
// ),
// 500000500000,
// i64
// );
// }
sum 1_000_000 0
"#
),
500000500000,
i64
);
}
// #[test]
// fn int_negate() {

954
compiler/gen_dev/tests/dev_str.rs Normal file
View file

@ -0,0 +1,954 @@
// #[macro_use]
// extern crate indoc;
#[macro_use]
mod helpers;
#[cfg(all(test, any(target_os = "linux", target_os = "macos"), any(target_arch = "x86_64"/*, target_arch = "aarch64"*/)))]
mod dev_str {
// use roc_std::{RocList, RocStr};
// #[test]
// fn str_split_bigger_delimiter_small_str() {
// assert_evals_to!(
// indoc!(
// r#"
// List.len (Str.split "hello" "JJJJ there")
// "#
// ),
// 1,
// i64
// );
// assert_evals_to!(
// indoc!(
// r#"
// when List.first (Str.split "JJJ" "JJJJ there") is
// Ok str ->
// Str.countGraphemes str
// _ ->
// -1
// "#
// ),
// 3,
// i64
// );
// }
// #[test]
// fn str_split_str_concat_repeated() {
// assert_evals_to!(
// indoc!(
// r#"
// when List.first (Str.split "JJJJJ" "JJJJ there") is
// Ok str ->
// str
// |> Str.concat str
// |> Str.concat str
// |> Str.concat str
// |> Str.concat str
// _ ->
// "Not Str!"
// "#
// ),
// RocStr::from_slice(b"JJJJJJJJJJJJJJJJJJJJJJJJJ"),
// RocStr
// );
// }
// #[test]
// fn str_split_small_str_bigger_delimiter() {
// assert_evals_to!(
// indoc!(
// r#"
// when
// List.first
// (Str.split "JJJ" "0123456789abcdefghi")
// is
// Ok str -> str
// _ -> ""
// "#
// ),
// RocStr::from_slice(b"JJJ"),
// RocStr
// );
// }
// #[test]
// fn str_split_big_str_small_delimiter() {
// assert_evals_to!(
// indoc!(
// r#"
// Str.split "01234567789abcdefghi?01234567789abcdefghi" "?"
// "#
// ),
// RocList::from_slice(&[
// RocStr::from_slice(b"01234567789abcdefghi"),
// RocStr::from_slice(b"01234567789abcdefghi")
// ]),
// RocList<RocStr>
// );
// assert_evals_to!(
// indoc!(
// r#"
// Str.split "01234567789abcdefghi 3ch 01234567789abcdefghi" "3ch"
// "#
// ),
// RocList::from_slice(&[
// RocStr::from_slice(b"01234567789abcdefghi "),
// RocStr::from_slice(b" 01234567789abcdefghi")
// ]),
// RocList<RocStr>
// );
// }
// #[test]
// fn str_split_small_str_small_delimiter() {
// assert_evals_to!(
// indoc!(
// r#"
// Str.split "J!J!J" "!"
// "#
// ),
// RocList::from_slice(&[
// RocStr::from_slice(b"J"),
// RocStr::from_slice(b"J"),
// RocStr::from_slice(b"J")
// ]),
// RocList<RocStr>
// );
// }
// #[test]
// fn str_split_bigger_delimiter_big_strs() {
// assert_evals_to!(
// indoc!(
// r#"
// Str.split
// "string to split is shorter"
// "than the delimiter which happens to be very very long"
// "#
// ),
// RocList::from_slice(&[RocStr::from_slice(b"string to split is shorter")]),
// RocList<RocStr>
// );
// }
// #[test]
// fn str_split_empty_strs() {
// assert_evals_to!(
// indoc!(
// r#"
// Str.split "" ""
// "#
// ),
// RocList::from_slice(&[RocStr::from_slice(b"")]),
// RocList<RocStr>
// );
// }
// #[test]
// fn str_split_minimal_example() {
// assert_evals_to!(
// indoc!(
// r#"
// Str.split "a," ","
// "#
// ),
// RocList::from_slice(&[RocStr::from_slice(b"a"), RocStr::from_slice(b"")]),
// RocList<RocStr>
// )
// }
// #[test]
// fn str_split_small_str_big_delimiter() {
// assert_evals_to!(
// indoc!(
// r#"
// Str.split
// "1---- ---- ---- ---- ----2---- ---- ---- ---- ----"
// "---- ---- ---- ---- ----"
// |> List.len
// "#
// ),
// 3,
// i64
// );
// assert_evals_to!(
// indoc!(
// r#"
// Str.split
// "1---- ---- ---- ---- ----2---- ---- ---- ---- ----"
// "---- ---- ---- ---- ----"
// "#
// ),
// RocList::from_slice(&[
// RocStr::from_slice(b"1"),
// RocStr::from_slice(b"2"),
// RocStr::from_slice(b"")
// ]),
// RocList<RocStr>
// );
// }
// #[test]
// fn str_split_small_str_20_char_delimiter() {
// assert_evals_to!(
// indoc!(
// r#"
// Str.split
// "3|-- -- -- -- -- -- |4|-- -- -- -- -- -- |"
// "|-- -- -- -- -- -- |"
// "#
// ),
// RocList::from_slice(&[
// RocStr::from_slice(b"3"),
// RocStr::from_slice(b"4"),
// RocStr::from_slice(b"")
// ]),
// RocList<RocStr>
// );
// }
// #[test]
// fn str_concat_big_to_big() {
// assert_evals_to!(
// indoc!(
// r#"
// Str.concat
// "First string that is fairly long. Longer strings make for different errors. "
// "Second string that is also fairly long. Two long strings test things that might not appear with short strings."
// "#
// ),
// RocStr::from_slice(b"First string that is fairly long. Longer strings make for different errors. Second string that is also fairly long. Two long strings test things that might not appear with short strings."),
// RocStr
// );
// }
#[test]
fn small_str_literal() {
assert_evals_to!(
"\"JJJJJJJJJJJJJJJ\"",
[
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0b1000_1111
],
[u8; 16]
);
}
// #[test]
// fn small_str_zeroed_literal() {
// // Verifies that we zero out unused bytes in the string.
// // This is important so that string equality tests don't randomly
// // fail due to unused memory being there!
// assert_evals_to!(
// "\"J\"",
// [
// 0x4a,
// 0x00,
// 0x00,
// 0x00,
// 0x00,
// 0x00,
// 0x00,
// 0x00,
// 0x00,
// 0x00,
// 0x00,
// 0x00,
// 0x00,
// 0x00,
// 0x00,
// 0b1000_0001
// ],
// [u8; 16]
// );
// }
#[test]
fn small_str_concat_empty_first_arg() {
assert_evals_to!(
r#"Str.concat "" "JJJJJJJJJJJJJJJ""#,
[
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0b1000_1111
],
[u8; 16]
);
}
#[test]
fn small_str_concat_empty_second_arg() {
assert_evals_to!(
r#"Str.concat "JJJJJJJJJJJJJJJ" """#,
[
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0b1000_1111
],
[u8; 16]
);
}
// #[test]
// fn small_str_concat_small_to_big() {
// assert_evals_to!(
// r#"Str.concat "abc" " this is longer than 15 chars""#,
// RocStr::from_slice(b"abc this is longer than 15 chars"),
// RocStr
// );
// }
#[test]
fn small_str_concat_small_to_small_staying_small() {
assert_evals_to!(
r#"Str.concat "J" "JJJJJJJJJJJJJJ""#,
[
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0x4a,
0b1000_1111
],
[u8; 16]
);
}
// #[test]
// fn small_str_concat_small_to_small_overflow_to_big() {
// assert_evals_to!(
// r#"Str.concat "abcdefghijklm" "nopqrstuvwxyz""#,
// RocStr::from_slice(b"abcdefghijklmnopqrstuvwxyz"),
// RocStr
// );
// }
// #[test]
// fn str_concat_empty() {
// assert_evals_to!(r#"Str.concat "" """#, RocStr::default(), RocStr);
// }
// #[test]
// fn small_str_is_empty() {
// assert_evals_to!(r#"Str.isEmpty "abc""#, false, bool);
// }
// #[test]
// fn big_str_is_empty() {
// assert_evals_to!(
// r#"Str.isEmpty "this is more than 15 chars long""#,
// false,
// bool
// );
// }
// #[test]
// fn empty_str_is_empty() {
// assert_evals_to!(r#"Str.isEmpty """#, true, bool);
// }
// #[test]
// fn str_starts_with() {
// assert_evals_to!(r#"Str.startsWith "hello world" "hell""#, true, bool);
// assert_evals_to!(r#"Str.startsWith "hello world" """#, true, bool);
// assert_evals_to!(r#"Str.startsWith "nope" "hello world""#, false, bool);
// assert_evals_to!(r#"Str.startsWith "hell" "hello world""#, false, bool);
// assert_evals_to!(r#"Str.startsWith "" "hello world""#, false, bool);
// }
// #[test]
// fn str_starts_with_code_point() {
// assert_evals_to!(
// &format!(r#"Str.startsWithCodePt "foobar" {}"#, 'f' as u32),
// true,
// bool
// );
// assert_evals_to!(
// &format!(r#"Str.startsWithCodePt "zoobar" {}"#, 'f' as u32),
// false,
// bool
// );
// }
// #[test]
// fn str_ends_with() {
// assert_evals_to!(r#"Str.endsWith "hello world" "world""#, true, bool);
// assert_evals_to!(r#"Str.endsWith "nope" "hello world""#, false, bool);
// assert_evals_to!(r#"Str.endsWith "" "hello world""#, false, bool);
// }
// #[test]
// fn str_count_graphemes_small_str() {
// assert_evals_to!(r#"Str.countGraphemes "å🤔""#, 2, usize);
// }
// #[test]
// fn str_count_graphemes_three_js() {
// assert_evals_to!(r#"Str.countGraphemes "JJJ""#, 3, usize);
// }
// #[test]
// fn str_count_graphemes_big_str() {
// assert_evals_to!(
// r#"Str.countGraphemes "6🤔å🤔e¥🤔çppkd🙃1jdal🦯asdfa∆ltråø˚waia8918.,🏅jjc""#,
// 45,
// usize
// );
// }
// #[test]
// fn str_starts_with_same_big_str() {
// assert_evals_to!(
// r#"Str.startsWith "123456789123456789" "123456789123456789""#,
// true,
// bool
// );
// }
// #[test]
// fn str_starts_with_different_big_str() {
// assert_evals_to!(
// r#"Str.startsWith "12345678912345678910" "123456789123456789""#,
// true,
// bool
// );
// }
// #[test]
// fn str_starts_with_same_small_str() {
// assert_evals_to!(r#"Str.startsWith "1234" "1234""#, true, bool);
// }
// #[test]
// fn str_starts_with_different_small_str() {
// assert_evals_to!(r#"Str.startsWith "1234" "12""#, true, bool);
// }
// #[test]
// fn str_starts_with_false_small_str() {
// assert_evals_to!(r#"Str.startsWith "1234" "23""#, false, bool);
// }
// #[test]
// fn str_from_int() {
// assert_evals_to!(
// r#"Str.fromInt 1234"#,
// roc_std::RocStr::from_slice("1234".as_bytes()),
// roc_std::RocStr
// );
// assert_evals_to!(
// r#"Str.fromInt 0"#,
// roc_std::RocStr::from_slice("0".as_bytes()),
// roc_std::RocStr
// );
// assert_evals_to!(
// r#"Str.fromInt -1"#,
// roc_std::RocStr::from_slice("-1".as_bytes()),
// roc_std::RocStr
// );
// let max = format!("{}", i64::MAX);
// assert_evals_to!(
// r#"Str.fromInt Num.maxInt"#,
// RocStr::from_slice(max.as_bytes()),
// RocStr
// );
// let min = format!("{}", i64::MIN);
// assert_evals_to!(
// r#"Str.fromInt Num.minInt"#,
// RocStr::from_slice(min.as_bytes()),
// RocStr
// );
// }
// #[test]
// fn str_from_utf8_pass_single_ascii() {
// assert_evals_to!(
// indoc!(
// r#"
// when Str.fromUtf8 [ 97 ] is
// Ok val -> val
// Err _ -> ""
// "#
// ),
// roc_std::RocStr::from_slice("a".as_bytes()),
// roc_std::RocStr
// );
// }
// #[test]
// fn str_from_utf8_pass_many_ascii() {
// assert_evals_to!(
// indoc!(
// r#"
// when Str.fromUtf8 [ 97, 98, 99, 0x7E ] is
// Ok val -> val
// Err _ -> ""
// "#
// ),
// roc_std::RocStr::from_slice("abc~".as_bytes()),
// roc_std::RocStr
// );
// }
// #[test]
// fn str_from_utf8_pass_single_unicode() {
// assert_evals_to!(
// indoc!(
// r#"
// when Str.fromUtf8 [ 0xE2, 0x88, 0x86 ] is
// Ok val -> val
// Err _ -> ""
// "#
// ),
// roc_std::RocStr::from_slice("∆".as_bytes()),
// roc_std::RocStr
// );
// }
// #[test]
// fn str_from_utf8_pass_many_unicode() {
// assert_evals_to!(
// indoc!(
// r#"
// when Str.fromUtf8 [ 0xE2, 0x88, 0x86, 0xC5, 0x93, 0xC2, 0xAC ] is
// Ok val -> val
// Err _ -> ""
// "#
// ),
// roc_std::RocStr::from_slice("∆œ¬".as_bytes()),
// roc_std::RocStr
// );
// }
// #[test]
// fn str_from_utf8_pass_single_grapheme() {
// assert_evals_to!(
// indoc!(
// r#"
// when Str.fromUtf8 [ 0xF0, 0x9F, 0x92, 0x96 ] is
// Ok val -> val
// Err _ -> ""
// "#
// ),
// roc_std::RocStr::from_slice("💖".as_bytes()),
// roc_std::RocStr
// );
// }
// #[test]
// fn str_from_utf8_pass_many_grapheme() {
// assert_evals_to!(
// indoc!(
// r#"
// when Str.fromUtf8 [ 0xF0, 0x9F, 0x92, 0x96, 0xF0, 0x9F, 0xA4, 0xA0, 0xF0, 0x9F, 0x9A, 0x80 ] is
// Ok val -> val
// Err _ -> ""
// "#
// ),
// roc_std::RocStr::from_slice("💖🤠🚀".as_bytes()),
// roc_std::RocStr
// );
// }
// #[test]
// fn str_from_utf8_pass_all() {
// assert_evals_to!(
// indoc!(
// r#"
// when Str.fromUtf8 [ 0xF0, 0x9F, 0x92, 0x96, 98, 0xE2, 0x88, 0x86 ] is
// Ok val -> val
// Err _ -> ""
// "#
// ),
// roc_std::RocStr::from_slice("💖b∆".as_bytes()),
// roc_std::RocStr
// );
// }
// #[test]
// fn str_from_utf8_fail_invalid_start_byte() {
// assert_evals_to!(
// indoc!(
// r#"
// when Str.fromUtf8 [ 97, 98, 0x80, 99 ] is
// Err (BadUtf8 InvalidStartByte byteIndex) ->
// if byteIndex == 2 then
// "a"
// else
// "b"
// _ -> ""
// "#
// ),
// roc_std::RocStr::from_slice("a".as_bytes()),
// roc_std::RocStr
// );
// }
// #[test]
// fn str_from_utf8_fail_unexpected_end_of_sequence() {
// assert_evals_to!(
// indoc!(
// r#"
// when Str.fromUtf8 [ 97, 98, 99, 0xC2 ] is
// Err (BadUtf8 UnexpectedEndOfSequence byteIndex) ->
// if byteIndex == 3 then
// "a"
// else
// "b"
// _ -> ""
// "#
// ),
// roc_std::RocStr::from_slice("a".as_bytes()),
// roc_std::RocStr
// );
// }
// #[test]
// fn str_from_utf8_fail_expected_continuation() {
// assert_evals_to!(
// indoc!(
// r#"
// when Str.fromUtf8 [ 97, 98, 99, 0xC2, 0x00 ] is
// Err (BadUtf8 ExpectedContinuation byteIndex) ->
// if byteIndex == 3 then
// "a"
// else
// "b"
// _ -> ""
// "#
// ),
// roc_std::RocStr::from_slice("a".as_bytes()),
// roc_std::RocStr
// );
// }
// #[test]
// fn str_from_utf8_fail_overlong_encoding() {
// assert_evals_to!(
// indoc!(
// r#"
// when Str.fromUtf8 [ 97, 0xF0, 0x80, 0x80, 0x80 ] is
// Err (BadUtf8 OverlongEncoding byteIndex) ->
// if byteIndex == 1 then
// "a"
// else
// "b"
// _ -> ""
// "#
// ),
// roc_std::RocStr::from_slice("a".as_bytes()),
// roc_std::RocStr
// );
// }
// #[test]
// fn str_from_utf8_fail_codepoint_too_large() {
// assert_evals_to!(
// indoc!(
// r#"
// when Str.fromUtf8 [ 97, 0xF4, 0x90, 0x80, 0x80 ] is
// Err (BadUtf8 CodepointTooLarge byteIndex) ->
// if byteIndex == 1 then
// "a"
// else
// "b"
// _ -> ""
// "#
// ),
// roc_std::RocStr::from_slice("a".as_bytes()),
// roc_std::RocStr
// );
// }
// #[test]
// fn str_from_utf8_fail_surrogate_half() {
// assert_evals_to!(
// indoc!(
// r#"
// when Str.fromUtf8 [ 97, 98, 0xED, 0xA0, 0x80 ] is
// Err (BadUtf8 EncodesSurrogateHalf byteIndex) ->
// if byteIndex == 2 then
// "a"
// else
// "b"
// _ -> ""
// "#
// ),
// roc_std::RocStr::from_slice("a".as_bytes()),
// roc_std::RocStr
// );
// }
// #[test]
// fn str_equality() {
// assert_evals_to!(r#""a" == "a""#, true, bool);
// assert_evals_to!(
// r#""loremipsumdolarsitamet" == "loremipsumdolarsitamet""#,
// true,
// bool
// );
// assert_evals_to!(r#""a" != "b""#, true, bool);
// assert_evals_to!(r#""a" == "b""#, false, bool);
// }
// #[test]
// fn str_clone() {
// use roc_std::RocStr;
// let long = RocStr::from_slice("loremipsumdolarsitamet".as_bytes());
// let short = RocStr::from_slice("x".as_bytes());
// let empty = RocStr::from_slice("".as_bytes());
// debug_assert_eq!(long.clone(), long);
// debug_assert_eq!(short.clone(), short);
// debug_assert_eq!(empty.clone(), empty);
// }
// #[test]
// fn nested_recursive_literal() {
// assert_evals_to!(
// indoc!(
// r#"
// Expr : [ Add Expr Expr, Val I64, Var I64 ]
// expr : Expr
// expr = Add (Add (Val 3) (Val 1)) (Add (Val 1) (Var 1))
// printExpr : Expr -> Str
// printExpr = \e ->
// when e is
// Add a b ->
// "Add ("
// |> Str.concat (printExpr a)
// |> Str.concat ") ("
// |> Str.concat (printExpr b)
// |> Str.concat ")"
// Val v -> "Val " |> Str.concat (Str.fromInt v)
// Var v -> "Var " |> Str.concat (Str.fromInt v)
// printExpr expr
// "#
// ),
// RocStr::from_slice(b"Add (Add (Val 3) (Val 1)) (Add (Val 1) (Var 1))"),
// RocStr
// );
// }
// #[test]
// fn str_join_comma_small() {
// assert_evals_to!(
// r#"Str.joinWith ["1", "2"] ", " "#,
// RocStr::from("1, 2"),
// RocStr
// );
// }
// #[test]
// fn str_join_comma_big() {
// assert_evals_to!(
// r#"Str.joinWith ["10000000", "2000000", "30000000"] ", " "#,
// RocStr::from("10000000, 2000000, 30000000"),
// RocStr
// );
// }
// #[test]
// fn str_join_comma_single() {
// assert_evals_to!(r#"Str.joinWith ["1"] ", " "#, RocStr::from("1"), RocStr);
// }
// #[test]
// fn str_from_float() {
// assert_evals_to!(r#"Str.fromFloat 3.14"#, RocStr::from("3.14"), RocStr);
// }
// #[test]
// fn str_to_utf8() {
// assert_evals_to!(
// r#"Str.toUtf8 "hello""#,
// RocList::from_slice(&[104, 101, 108, 108, 111]),
// RocList<u8>
// );
// assert_evals_to!(
// r#"Str.toUtf8 "this is a long string""#,
// RocList::from_slice(&[
// 116, 104, 105, 115, 32, 105, 115, 32, 97, 32, 108, 111, 110, 103, 32, 115, 116,
// 114, 105, 110, 103
// ]),
// RocList<u8>
// );
// }
// #[test]
// fn str_from_utf8_range() {
// assert_evals_to!(
// indoc!(
// r#"
// bytes = Str.toUtf8 "hello"
// when Str.fromUtf8Range bytes { count: 5, start: 0 } is
// Ok utf8String -> utf8String
// _ -> ""
// "#
// ),
// RocStr::from("hello"),
// RocStr
// );
// }
// #[test]
// fn str_from_utf8_range_slice() {
// assert_evals_to!(
// indoc!(
// r#"
// bytes = Str.toUtf8 "hello"
// when Str.fromUtf8Range bytes { count: 4, start: 1 } is
// Ok utf8String -> utf8String
// _ -> ""
// "#
// ),
// RocStr::from("ello"),
// RocStr
// );
// }
// #[test]
// fn str_from_utf8_range_slice_not_end() {
// assert_evals_to!(
// indoc!(
// r#"
// bytes = Str.toUtf8 "hello"
// when Str.fromUtf8Range bytes { count: 3, start: 1 } is
// Ok utf8String -> utf8String
// _ -> ""
// "#
// ),
// RocStr::from("ell"),
// RocStr
// );
// }
// #[test]
// fn str_from_utf8_range_order_does_not_matter() {
// assert_evals_to!(
// indoc!(
// r#"
// bytes = Str.toUtf8 "hello"
// when Str.fromUtf8Range bytes { start: 1, count: 3 } is
// Ok utf8String -> utf8String
// _ -> ""
// "#
// ),
// RocStr::from("ell"),
// RocStr
// );
// }
// #[test]
// fn str_from_utf8_range_out_of_bounds_start_value() {
// assert_evals_to!(
// indoc!(
// r#"
// bytes = Str.toUtf8 "hello"
// when Str.fromUtf8Range bytes { start: 7, count: 3 } is
// Ok _ -> ""
// Err (BadUtf8 _ _) -> ""
// Err OutOfBounds -> "out of bounds"
// "#
// ),
// RocStr::from("out of bounds"),
// RocStr
// );
// }
// #[test]
// fn str_from_utf8_range_count_too_high() {
// assert_evals_to!(
// indoc!(
// r#"
// bytes = Str.toUtf8 "hello"
// when Str.fromUtf8Range bytes { start: 0, count: 6 } is
// Ok _ -> ""
// Err (BadUtf8 _ _) -> ""
// Err OutOfBounds -> "out of bounds"
// "#
// ),
// RocStr::from("out of bounds"),
// RocStr
// );
// }
// #[test]
// fn str_from_utf8_range_count_too_high_for_start() {
// assert_evals_to!(
// indoc!(
// r#"
// bytes = Str.toUtf8 "hello"
// when Str.fromUtf8Range bytes { start: 4, count: 3 } is
// Ok _ -> ""
// Err (BadUtf8 _ _) -> ""
// Err OutOfBounds -> "out of bounds"
// "#
// ),
// RocStr::from("out of bounds"),
// RocStr
// );
// }
}

13
compiler/gen_dev/tests/helpers/eval.rs
View file

@ -94,10 +94,12 @@ pub fn helper<'a>(
let main_fn_layout = loaded.entry_point.layout;
let mut layout_ids = roc_mono::layout::LayoutIds::default();
let main_fn_name = layout_ids
let main_fn_name_base = layout_ids
.get_toplevel(main_fn_symbol, &main_fn_layout)
.to_symbol_string(main_fn_symbol, &interns);
let main_fn_name = format!("roc_{}_exposed", main_fn_name_base);
let mut lines = Vec::new();
// errors whose reporting we delay (so we can see that code gen generates runtime errors)
let mut delayed_errors = Vec::new();
@ -143,12 +145,13 @@ pub fn helper<'a>(
}
for problem in type_problems {
let report = type_problem(&alloc, module_path.clone(), problem);
let mut buf = String::new();
if let Some(report) = type_problem(&alloc, module_path.clone(), problem) {
let mut buf = String::new();
report.render_color_terminal(&mut buf, &alloc, &palette);
report.render_color_terminal(&mut buf, &alloc, &palette);
lines.push(buf);
lines.push(buf);
}
}
for problem in mono_problems {

12
compiler/gen_llvm/src/llvm/bitcode.rs
View file

@ -10,7 +10,7 @@ use inkwell::types::{BasicType, BasicTypeEnum};
use inkwell::values::{BasicValue, BasicValueEnum, CallSiteValue, FunctionValue, InstructionValue};
use inkwell::AddressSpace;
use roc_module::symbol::Symbol;
use roc_mono::layout::{Layout, LayoutIds, UnionLayout};
use roc_mono::layout::{LambdaSet, Layout, LayoutIds, UnionLayout};
pub fn call_bitcode_fn<'a, 'ctx, 'env>(
env: &Env<'a, 'ctx, 'env>,
@ -189,7 +189,7 @@ fn build_has_tag_id_help<'a, 'ctx, 'env>(
pub fn build_transform_caller<'a, 'ctx, 'env>(
env: &Env<'a, 'ctx, 'env>,
function: FunctionValue<'ctx>,
closure_data_layout: Layout<'a>,
closure_data_layout: LambdaSet<'a>,
argument_layouts: &[Layout<'a>],
) -> FunctionValue<'ctx> {
let fn_name: &str = &format!(
@ -212,7 +212,7 @@ pub fn build_transform_caller<'a, 'ctx, 'env>(
fn build_transform_caller_help<'a, 'ctx, 'env>(
env: &Env<'a, 'ctx, 'env>,
roc_function: FunctionValue<'ctx>,
closure_data_layout: Layout<'a>,
closure_data_layout: LambdaSet<'a>,
argument_layouts: &[Layout<'a>],
fn_name: &str,
) -> FunctionValue<'ctx> {
@ -270,7 +270,7 @@ fn build_transform_caller_help<'a, 'ctx, 'env>(
arguments_cast.push(argument);
}
match closure_data_layout {
match closure_data_layout.runtime_representation() {
Layout::Struct(&[]) => {
// nothing to add
}
@ -529,7 +529,7 @@ pub fn build_eq_wrapper<'a, 'ctx, 'env>(
pub fn build_compare_wrapper<'a, 'ctx, 'env>(
env: &Env<'a, 'ctx, 'env>,
roc_function: FunctionValue<'ctx>,
closure_data_layout: Layout<'a>,
closure_data_layout: LambdaSet<'a>,
layout: &Layout<'a>,
) -> FunctionValue<'ctx> {
let block = env.builder.get_insert_block().expect("to be in a function");
@ -595,7 +595,7 @@ pub fn build_compare_wrapper<'a, 'ctx, 'env>(
let default = [value1, value2];
let arguments_cast = match closure_data_layout {
let arguments_cast = match closure_data_layout.runtime_representation() {
Layout::Struct(&[]) => {
// nothing to add
&default

340
compiler/gen_llvm/src/llvm/build.rs
View file

@ -15,7 +15,7 @@ use crate::llvm::build_list::{
};
use crate::llvm::build_str::{
empty_str, str_concat, str_count_graphemes, str_ends_with, str_from_float, str_from_int,
str_from_utf8, str_from_utf8_range, str_join_with, str_number_of_bytes, str_split,
str_from_utf8, str_from_utf8_range, str_join_with, str_number_of_bytes, str_repeat, str_split,
str_starts_with, str_starts_with_code_point, str_to_utf8,
};
use crate::llvm::compare::{generic_eq, generic_neq};
@ -638,7 +638,7 @@ pub fn construct_optimization_passes<'a>(
let pmb = PassManagerBuilder::create();
match opt_level {
OptLevel::Normal => {
OptLevel::Development | OptLevel::Normal => {
pmb.set_optimization_level(OptimizationLevel::None);
}
OptLevel::Optimize => {
@ -704,8 +704,14 @@ fn promote_to_main_function<'a, 'ctx, 'env>(
let main_fn_name = "$Test.main";
// Add main to the module.
let main_fn =
expose_function_to_host_help_c_abi(env, main_fn_name, roc_main_fn, &[], main_fn_name);
let main_fn = expose_function_to_host_help_c_abi(
env,
main_fn_name,
roc_main_fn,
&[],
top_level.result,
main_fn_name,
);
(main_fn_name, main_fn)
}
@ -1164,8 +1170,16 @@ pub fn build_exp_expr<'a, 'ctx, 'env>(
StructAtIndex {
index, structure, ..
} => {
let (value, layout) = load_symbol_and_layout(scope, structure);
let layout = if let Layout::LambdaSet(lambda_set) = layout {
lambda_set.runtime_representation()
} else {
*layout
};
// extract field from a record
match load_symbol_and_layout(scope, structure) {
match (value, layout) {
(StructValue(argument), Layout::Struct(fields)) => {
debug_assert!(!fields.is_empty());
env.builder
@ -2601,6 +2615,18 @@ pub fn load_symbol_and_layout<'a, 'ctx, 'b>(
None => panic!("There was no entry for {:?} in scope {:?}", symbol, scope),
}
}
pub fn load_symbol_and_lambda_set<'a, 'ctx, 'b>(
scope: &'b Scope<'a, 'ctx>,
symbol: &Symbol,
) -> (BasicValueEnum<'ctx>, LambdaSet<'a>) {
match scope.get(symbol) {
Some((Layout::LambdaSet(lambda_set), ptr)) => (*ptr, *lambda_set),
Some((other, ptr)) => panic!("Not a lambda set: {:?}, {:?}", other, ptr),
None => panic!("There was no entry for {:?} in scope {:?}", symbol, scope),
}
}
fn access_index_struct_value<'ctx>(
builder: &Builder<'ctx>,
from_value: StructValue<'ctx>,
@ -3055,6 +3081,7 @@ fn expose_function_to_host<'a, 'ctx, 'env>(
symbol: Symbol,
roc_function: FunctionValue<'ctx>,
arguments: &[Layout<'a>],
return_layout: Layout<'a>,
) {
// Assumption: there is only one specialization of a host-exposed function
let ident_string = symbol.as_str(&env.interns);
@ -3065,26 +3092,19 @@ fn expose_function_to_host<'a, 'ctx, 'env>(
ident_string,
roc_function,
arguments,
return_layout,
&c_function_name,
);
}
fn expose_function_to_host_help_c_abi<'a, 'ctx, 'env>(
fn expose_function_to_host_help_c_abi_generic<'a, 'ctx, 'env>(
env: &Env<'a, 'ctx, 'env>,
ident_string: &str,
roc_function: FunctionValue<'ctx>,
arguments: &[Layout<'a>],
c_function_name: &str,
) -> FunctionValue<'ctx> {
let context = env.context;
let wrapper_return_type = context.struct_type(
&[
context.i64_type().into(),
roc_function.get_type().get_return_type().unwrap(),
],
false,
);
// NOTE we ingore env.is_gen_test here
let wrapper_return_type = roc_function.get_type().get_return_type().unwrap();
let mut cc_argument_types = Vec::with_capacity_in(arguments.len(), env.arena);
for layout in arguments {
@ -3095,6 +3115,7 @@ fn expose_function_to_host_help_c_abi<'a, 'ctx, 'env>(
// let mut argument_types = roc_function.get_type().get_param_types();
let mut argument_types = cc_argument_types;
let return_type = wrapper_return_type;
let output_type = return_type.ptr_type(AddressSpace::Generic);
argument_types.push(output_type.into());
@ -3122,9 +3143,11 @@ fn expose_function_to_host_help_c_abi<'a, 'ctx, 'env>(
debug_info_init!(env, c_function);
// drop the final argument, which is the pointer we write the result into
let args = c_function.get_params();
let output_arg_index = args.len() - 1;
let args = &args[..args.len() - 1];
let args_vector = c_function.get_params();
let mut args = args_vector.as_slice();
let args_length = args.len();
args = &args[..args.len() - 1];
let mut arguments_for_call = Vec::with_capacity_in(args.len(), env.arena);
@ -3169,19 +3192,187 @@ fn expose_function_to_host_help_c_abi<'a, 'ctx, 'env>(
let call_unwrapped_result = call_unwrapped.try_as_basic_value().left().unwrap();
make_good_roc_result(env, call_unwrapped_result)
// make_good_roc_result(env, call_unwrapped_result)
call_unwrapped_result
}
};
let output_arg_index = args_length - 1;
let output_arg = c_function
.get_nth_param(output_arg_index as u32)
.unwrap()
.into_pointer_value();
builder.build_store(output_arg, call_result);
builder.build_return(None);
c_function
}
fn expose_function_to_host_help_c_abi<'a, 'ctx, 'env>(
env: &Env<'a, 'ctx, 'env>,
ident_string: &str,
roc_function: FunctionValue<'ctx>,
arguments: &[Layout<'a>],
return_layout: Layout<'a>,
c_function_name: &str,
) -> FunctionValue<'ctx> {
let context = env.context;
// a generic version that writes the result into a passed *u8 pointer
if !env.is_gen_test {
expose_function_to_host_help_c_abi_generic(
env,
roc_function,
arguments,
&format!("{}_generic", c_function_name),
);
}
let wrapper_return_type = if env.is_gen_test {
context
.struct_type(
&[
context.i64_type().into(),
roc_function.get_type().get_return_type().unwrap(),
],
false,
)
.into()
} else {
roc_function.get_type().get_return_type().unwrap()
};
let mut cc_argument_types = Vec::with_capacity_in(arguments.len(), env.arena);
for layout in arguments {
cc_argument_types.push(to_cc_type(env, layout));
}
// STEP 1: turn `f : a,b,c -> d` into `f : a,b,c, &d -> {}` if the C abi demands it
let mut argument_types = cc_argument_types;
let return_type = wrapper_return_type;
let cc_return = to_cc_return(env, &return_layout);
let c_function_type = match cc_return {
CCReturn::Void if !env.is_gen_test => {
env.context.void_type().fn_type(&argument_types, false)
}
CCReturn::Return if !env.is_gen_test => return_type.fn_type(&argument_types, false),
_ => {
let output_type = return_type.ptr_type(AddressSpace::Generic);
argument_types.push(output_type.into());
env.context.void_type().fn_type(&argument_types, false)
}
};
let c_function = add_func(
env.module,
c_function_name,
c_function_type,
Linkage::External,
C_CALL_CONV,
);
let subprogram = env.new_subprogram(c_function_name);
c_function.set_subprogram(subprogram);
// STEP 2: build the exposed function's body
let builder = env.builder;
let context = env.context;
let entry = context.append_basic_block(c_function, "entry");
builder.position_at_end(entry);
debug_info_init!(env, c_function);
// drop the final argument, which is the pointer we write the result into
let args_vector = c_function.get_params();
let mut args = args_vector.as_slice();
let args_length = args.len();
match cc_return {
CCReturn::Return if !env.is_gen_test => {
debug_assert_eq!(args.len(), roc_function.get_params().len());
}
CCReturn::Void if !env.is_gen_test => {
debug_assert_eq!(args.len(), roc_function.get_params().len());
}
_ => {
args = &args[..args.len() - 1];
debug_assert_eq!(args.len(), roc_function.get_params().len());
}
}
let mut arguments_for_call = Vec::with_capacity_in(args.len(), env.arena);
let it = args.iter().zip(roc_function.get_type().get_param_types());
for (arg, fastcc_type) in it {
let arg_type = arg.get_type();
if arg_type == fastcc_type {
// the C and Fast calling conventions agree
arguments_for_call.push(*arg);
} else {
let cast = complex_bitcast_check_size(env, *arg, fastcc_type, "to_fastcc_type");
arguments_for_call.push(cast);
}
}
let arguments_for_call = &arguments_for_call.into_bump_slice();
let call_result = {
if env.is_gen_test {
let roc_wrapper_function = make_exception_catcher(env, roc_function);
debug_assert_eq!(
arguments_for_call.len(),
roc_wrapper_function.get_params().len()
);
builder.position_at_end(entry);
let call_wrapped = builder.build_call(
roc_wrapper_function,
arguments_for_call,
"call_wrapped_function",
);
call_wrapped.set_call_convention(FAST_CALL_CONV);
call_wrapped.try_as_basic_value().left().unwrap()
} else {
let call_unwrapped =
builder.build_call(roc_function, arguments_for_call, "call_unwrapped_function");
call_unwrapped.set_call_convention(FAST_CALL_CONV);
let call_unwrapped_result = call_unwrapped.try_as_basic_value().left().unwrap();
// make_good_roc_result(env, call_unwrapped_result)
call_unwrapped_result
}
};
match cc_return {
CCReturn::Void if !env.is_gen_test => {
// TODO return empty struct here?
builder.build_return(None);
}
CCReturn::Return if !env.is_gen_test => {
builder.build_return(Some(&call_result));
}
_ => {
let output_arg_index = args_length - 1;
let output_arg = c_function
.get_nth_param(output_arg_index as u32)
.unwrap()
.into_pointer_value();
builder.build_store(output_arg, call_result);
builder.build_return(None);
}
}
// STEP 3: build a {} -> u64 function that gives the size of the return type
let size_function_type = env.context.i64_type().fn_type(&[], false);
let size_function_name: String = format!("roc__{}_size", ident_string);
@ -3695,7 +3886,13 @@ fn build_proc_header<'a, 'ctx, 'env>(
if env.exposed_to_host.contains(&symbol) {
let arguments = Vec::from_iter_in(proc.args.iter().map(|(layout, _)| *layout), env.arena);
expose_function_to_host(env, symbol, fn_val, arguments.into_bump_slice());
expose_function_to_host(
env,
symbol,
fn_val,
arguments.into_bump_slice(),
proc.ret_layout,
);
}
fn_val
@ -3762,23 +3959,17 @@ pub fn build_closure_caller<'a, 'ctx, 'env>(
builder.position_at_end(entry);
let mut parameters = function_value.get_params();
let output = parameters.pop().unwrap().into_pointer_value();
let mut evaluator_arguments = function_value.get_params();
let closure_data = if let Some(closure_data_ptr) = parameters.pop() {
let closure_data =
builder.build_load(closure_data_ptr.into_pointer_value(), "load_closure_data");
// the final parameter is the output pointer, pop it
let output = evaluator_arguments.pop().unwrap().into_pointer_value();
env.arena.alloc([closure_data]) as &[_]
} else {
&[]
};
let mut parameters = parameters;
for param in parameters.iter_mut() {
debug_assert!(param.is_pointer_value());
*param = builder.build_load(param.into_pointer_value(), "load_param");
// NOTE this may be incorrect in the long run
// here we load any argument that is a pointer
for param in evaluator_arguments.iter_mut() {
if param.is_pointer_value() {
*param = builder.build_load(param.into_pointer_value(), "load_param");
}
}
let call_result = if env.is_gen_test {
@ -3787,13 +3978,13 @@ pub fn build_closure_caller<'a, 'ctx, 'env>(
function_value,
evaluator,
evaluator.get_call_conventions(),
closure_data,
&evaluator_arguments,
result_type,
)
} else {
let call = env
.builder
.build_call(evaluator, closure_data, "call_function");
.build_call(evaluator, &evaluator_arguments, "call_function");
call.set_call_convention(evaluator.get_call_conventions());
@ -4094,7 +4285,7 @@ fn roc_function_call<'a, 'ctx, 'env>(
layout_ids: &mut LayoutIds<'a>,
transform: FunctionValue<'ctx>,
closure_data: BasicValueEnum<'ctx>,
closure_data_layout: Layout<'a>,
lambda_set: LambdaSet<'a>,
closure_data_is_owned: bool,
argument_layouts: &[Layout<'a>],
) -> RocFunctionCall<'ctx> {
@ -4105,15 +4296,15 @@ fn roc_function_call<'a, 'ctx, 'env>(
.build_alloca(closure_data.get_type(), "closure_data_ptr");
env.builder.build_store(closure_data_ptr, closure_data);
let stepper_caller =
build_transform_caller(env, transform, closure_data_layout, argument_layouts)
.as_global_value()
.as_pointer_value();
let inc_closure_data = build_inc_n_wrapper(env, layout_ids, &closure_data_layout)
let stepper_caller = build_transform_caller(env, transform, lambda_set, argument_layouts)
.as_global_value()
.as_pointer_value();
let inc_closure_data =
build_inc_n_wrapper(env, layout_ids, &lambda_set.runtime_representation())
.as_global_value()
.as_pointer_value();
let closure_data_is_owned = env
.context
.bool_type()
@ -4167,7 +4358,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
let function = passed_function_at_index!(2);
let (closure, closure_layout) = load_symbol_and_layout(scope, &args[3]);
let (closure, closure_layout) = load_symbol_and_lambda_set(scope, &args[3]);
match list_layout {
Layout::Builtin(Builtin::EmptyList) => default,
@ -4179,7 +4370,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
layout_ids,
function,
closure,
*closure_layout,
closure_layout,
function_owns_closure_data,
argument_layouts,
);
@ -4209,7 +4400,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
let function = passed_function_at_index!(1);
let (closure, closure_layout) = load_symbol_and_layout(scope, &args[2]);
let (closure, closure_layout) = load_symbol_and_lambda_set(scope, &args[2]);
match (list_layout, return_layout) {
(Layout::Builtin(Builtin::EmptyList), _) => empty_list(env),
@ -4224,7 +4415,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
layout_ids,
function,
closure,
*closure_layout,
closure_layout,
function_owns_closure_data,
argument_layouts,
);
@ -4241,7 +4432,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
let (list2, list2_layout) = load_symbol_and_layout(scope, &args[1]);
let function = passed_function_at_index!(2);
let (closure, closure_layout) = load_symbol_and_layout(scope, &args[3]);
let (closure, closure_layout) = load_symbol_and_lambda_set(scope, &args[3]);
match (list1_layout, list2_layout, return_layout) {
(
@ -4256,7 +4447,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
layout_ids,
function,
closure,
*closure_layout,
closure_layout,
function_owns_closure_data,
argument_layouts,
);
@ -4285,7 +4476,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
let (list3, list3_layout) = load_symbol_and_layout(scope, &args[2]);
let function = passed_function_at_index!(3);
let (closure, closure_layout) = load_symbol_and_layout(scope, &args[4]);
let (closure, closure_layout) = load_symbol_and_lambda_set(scope, &args[4]);
match (list1_layout, list2_layout, list3_layout, return_layout) {
(
@ -4302,7 +4493,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
layout_ids,
function,
closure,
*closure_layout,
closure_layout,
function_owns_closure_data,
argument_layouts,
);
@ -4334,7 +4525,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
let function = passed_function_at_index!(1);
let (closure, closure_layout) = load_symbol_and_layout(scope, &args[2]);
let (closure, closure_layout) = load_symbol_and_lambda_set(scope, &args[2]);
match (list_layout, return_layout) {
(Layout::Builtin(Builtin::EmptyList), _) => empty_list(env),
@ -4349,7 +4540,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
layout_ids,
function,
closure,
*closure_layout,
closure_layout,
function_owns_closure_data,
argument_layouts,
);
@ -4367,7 +4558,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
let function = passed_function_at_index!(1);
let (closure, closure_layout) = load_symbol_and_layout(scope, &args[2]);
let (closure, closure_layout) = load_symbol_and_lambda_set(scope, &args[2]);
match list_layout {
Layout::Builtin(Builtin::EmptyList) => empty_list(env),
@ -4379,7 +4570,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
layout_ids,
function,
closure,
*closure_layout,
closure_layout,
function_owns_closure_data,
argument_layouts,
);
@ -4397,7 +4588,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
let function = passed_function_at_index!(1);
let (closure, closure_layout) = load_symbol_and_layout(scope, &args[2]);
let (closure, closure_layout) = load_symbol_and_lambda_set(scope, &args[2]);
match (list_layout, return_layout) {
(_, Layout::Builtin(Builtin::EmptyList))
@ -4413,7 +4604,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
layout_ids,
function,
closure,
*closure_layout,
closure_layout,
function_owns_closure_data,
argument_layouts,
);
@ -4441,7 +4632,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
let function = passed_function_at_index!(1);
let (closure, closure_layout) = load_symbol_and_layout(scope, &args[2]);
let (closure, closure_layout) = load_symbol_and_lambda_set(scope, &args[2]);
match (list_layout, return_layout) {
(_, Layout::Builtin(Builtin::EmptyList))
@ -4457,7 +4648,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
layout_ids,
function,
closure,
*closure_layout,
closure_layout,
function_owns_closure_data,
argument_layouts,
);
@ -4494,7 +4685,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
let function = passed_function_at_index!(1);
let (closure, closure_layout) = load_symbol_and_layout(scope, &args[2]);
let (closure, closure_layout) = load_symbol_and_lambda_set(scope, &args[2]);
match list_layout {
Layout::Builtin(Builtin::EmptyList) => empty_list(env),
@ -4504,7 +4695,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
let argument_layouts = &[**element_layout, **element_layout];
let compare_wrapper =
build_compare_wrapper(env, function, *closure_layout, element_layout)
build_compare_wrapper(env, function, closure_layout, element_layout)
.as_global_value()
.as_pointer_value();
@ -4513,7 +4704,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
layout_ids,
function,
closure,
*closure_layout,
closure_layout,
function_owns_closure_data,
argument_layouts,
);
@ -4535,7 +4726,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
let (dict, dict_layout) = load_symbol_and_layout(scope, &args[0]);
let (default, default_layout) = load_symbol_and_layout(scope, &args[1]);
let function = passed_function_at_index!(2);
let (closure, closure_layout) = load_symbol_and_layout(scope, &args[3]);
let (closure, closure_layout) = load_symbol_and_lambda_set(scope, &args[3]);
match dict_layout {
Layout::Builtin(Builtin::EmptyDict) => {
@ -4550,7 +4741,7 @@ fn run_higher_order_low_level<'a, 'ctx, 'env>(
layout_ids,
function,
closure,
*closure_layout,
closure_layout,
function_owns_closure_data,
argument_layouts,
);
@ -4662,6 +4853,12 @@ fn run_low_level<'a, 'ctx, 'env>(
str_to_utf8(env, string.into_struct_value())
}
StrRepeat => {
// Str.repeat : Str, Nat -> Str
debug_assert_eq!(args.len(), 2);
str_repeat(env, scope, args[0], args[1])
}
StrSplit => {
// Str.split : Str, Str -> List Str
debug_assert_eq!(args.len(), 2);
@ -4836,7 +5033,7 @@ fn run_low_level<'a, 'ctx, 'env>(
Usize | Int128 | Int64 | Int32 | Int16 | Int8 => {
build_int_unary_op(env, arg.into_int_value(), arg_builtin, op)
}
Float128 | Float64 | Float32 | Float16 => {
Float128 | Float64 | Float32 => {
build_float_unary_op(env, arg.into_float_value(), op)
}
_ => {
@ -4932,7 +5129,7 @@ fn run_low_level<'a, 'ctx, 'env>(
"lt_or_gt",
)
}
Float128 | Float64 | Float32 | Float16 => {
Float128 | Float64 | Float32 => {
let are_equal = env.builder.build_float_compare(
FloatPredicate::OEQ,
lhs_arg.into_float_value(),
@ -5378,8 +5575,7 @@ fn to_cc_type_builtin<'a, 'ctx, 'env>(
| Builtin::Decimal
| Builtin::Float128
| Builtin::Float64
| Builtin::Float32
| Builtin::Float16 => basic_type_from_builtin(env, builtin),
| Builtin::Float32 => basic_type_from_builtin(env, builtin),
Builtin::Str | Builtin::EmptyStr | Builtin::List(_) | Builtin::EmptyList => {
env.str_list_c_abi().into()
}
@ -5742,7 +5938,7 @@ pub fn build_num_binop<'a, 'ctx, 'env>(
rhs_layout,
op,
),
Float128 | Float64 | Float32 | Float16 => build_float_binop(
Float128 | Float64 | Float32 => build_float_binop(
env,
parent,
lhs_arg.into_float_value(),

10
compiler/gen_llvm/src/llvm/build_hash.rs
View file

@ -59,6 +59,15 @@ fn build_hash_layout<'a, 'ctx, 'env>(
val.into_struct_value(),
),
Layout::LambdaSet(lambda_set) => build_hash_layout(
env,
layout_ids,
seed,
val,
&lambda_set.runtime_representation(),
when_recursive,
),
Layout::Union(union_layout) => {
build_hash_tag(env, layout_ids, layout, union_layout, seed, val)
}
@ -123,7 +132,6 @@ fn hash_builtin<'a, 'ctx, 'env>(
| Builtin::Float64
| Builtin::Float32
| Builtin::Float128
| Builtin::Float16
| Builtin::Decimal
| Builtin::Usize => {
let hash_bytes = store_and_use_as_u8_ptr(env, val, layout);

47
compiler/gen_llvm/src/llvm/build_str.rs
View file

@ -1,5 +1,5 @@
use crate::llvm::bitcode::{call_bitcode_fn, call_void_bitcode_fn};
use crate::llvm::build::{complex_bitcast, struct_from_fields, Env, Scope};
use crate::llvm::build::{complex_bitcast, Env, Scope};
use crate::llvm::build_list::{allocate_list, call_bitcode_fn_returns_list, store_list};
use inkwell::builder::Builder;
use inkwell::values::{BasicValueEnum, FunctionValue, IntValue, PointerValue, StructValue};
@ -12,6 +12,18 @@ use super::build::load_symbol;
pub static CHAR_LAYOUT: Layout = Layout::Builtin(Builtin::Int8);
/// Str.repeat : Str, Nat -> Str
pub fn str_repeat<'a, 'ctx, 'env>(
env: &Env<'a, 'ctx, 'env>,
scope: &Scope<'a, 'ctx>,
str_symbol: Symbol,
count_symbol: Symbol,
) -> BasicValueEnum<'ctx> {
let str_c_abi = str_symbol_to_c_abi(env, scope, str_symbol);
let count = load_symbol(scope, &count_symbol);
call_bitcode_fn(env, &[str_c_abi.into(), count], bitcode::STR_REPEAT)
}
/// Str.split : Str, Str -> List Str
pub fn str_split<'a, 'ctx, 'env>(
env: &Env<'a, 'ctx, 'env>,
@ -268,48 +280,19 @@ fn decode_from_utf8_result<'a, 'ctx, 'env>(
let ctx = env.context;
let fields = match env.ptr_bytes {
8 => [
8 | 4 => [
env.ptr_int().into(),
super::convert::zig_str_type(env).into(),
env.context.bool_type().into(),
ctx.i8_type().into(),
],
4 => [
super::convert::zig_str_type(env).into(),
env.ptr_int().into(),
env.context.bool_type().into(),
ctx.i8_type().into(),
],
_ => unreachable!(),
};
let record_type = env.context.struct_type(&fields, false);
match env.ptr_bytes {
8 => {
let zig_struct = builder
.build_load(pointer, "load_utf8_validate_bytes_result")
.into_struct_value();
let string = builder
.build_extract_value(zig_struct, 0, "string")
.unwrap();
let byte_index = builder
.build_extract_value(zig_struct, 1, "byte_index")
.unwrap();
let is_ok = builder.build_extract_value(zig_struct, 2, "is_ok").unwrap();
let problem_code = builder
.build_extract_value(zig_struct, 3, "problem_code")
.unwrap();
let values = [byte_index, string, is_ok, problem_code];
struct_from_fields(env, record_type, values.iter().copied().enumerate())
}
4 => {
8 | 4 => {
let result_ptr_cast = env
.builder
.build_bitcast(

5
compiler/gen_llvm/src/llvm/compare.rs
View file

@ -103,7 +103,6 @@ fn build_eq_builtin<'a, 'ctx, 'env>(
Builtin::Float128 => float_cmp(FloatPredicate::OEQ, "eq_f128"),
Builtin::Float64 => float_cmp(FloatPredicate::OEQ, "eq_f64"),
Builtin::Float32 => float_cmp(FloatPredicate::OEQ, "eq_f32"),
Builtin::Float16 => float_cmp(FloatPredicate::OEQ, "eq_f16"),
Builtin::Str => str_equal(env, lhs_val, rhs_val),
Builtin::List(elem) => build_list_eq(
@ -156,6 +155,8 @@ fn build_eq<'a, 'ctx, 'env>(
rhs_val.into_struct_value(),
),
Layout::LambdaSet(_) => unreachable!("cannot compare closures"),
Layout::Union(union_layout) => build_tag_eq(
env,
layout_ids,
@ -245,7 +246,6 @@ fn build_neq_builtin<'a, 'ctx, 'env>(
Builtin::Float128 => float_cmp(FloatPredicate::ONE, "neq_f128"),
Builtin::Float64 => float_cmp(FloatPredicate::ONE, "neq_f64"),
Builtin::Float32 => float_cmp(FloatPredicate::ONE, "neq_f32"),
Builtin::Float16 => float_cmp(FloatPredicate::ONE, "neq_f16"),
Builtin::Str => {
let is_equal = str_equal(env, lhs_val, rhs_val).into_int_value();
@ -336,6 +336,7 @@ fn build_neq<'a, 'ctx, 'env>(
Layout::RecursivePointer => {
unreachable!("recursion pointers should never be compared directly")
}
Layout::LambdaSet(_) => unreachable!("cannot compare closure"),
}
}

2
compiler/gen_llvm/src/llvm/convert.rs
View file

@ -27,6 +27,7 @@ pub fn basic_type_from_layout<'a, 'ctx, 'env>(
match layout {
Struct(sorted_fields) => basic_type_from_record(env, sorted_fields),
LambdaSet(lambda_set) => basic_type_from_layout(env, &lambda_set.runtime_representation()),
Union(union_layout) => {
use UnionLayout::*;
@ -96,7 +97,6 @@ pub fn basic_type_from_builtin<'a, 'ctx, 'env>(
Float128 => context.f128_type().as_basic_type_enum(),
Float64 => context.f64_type().as_basic_type_enum(),
Float32 => context.f32_type().as_basic_type_enum(),
Float16 => context.f16_type().as_basic_type_enum(),
Dict(_, _) | EmptyDict => zig_dict_type(env).into(),
Set(_) | EmptySet => zig_dict_type(env).into(),
List(_) | EmptyList => zig_list_type(env).into(),

8
compiler/gen_llvm/src/llvm/refcounting.rs
View file

@ -626,6 +626,14 @@ fn modify_refcount_layout_build_function<'a, 'ctx, 'env>(
Some(function)
}
},
LambdaSet(lambda_set) => modify_refcount_layout_build_function(
env,
parent,
layout_ids,
mode,
when_recursive,
&lambda_set.runtime_representation(),
),
}
}

107
compiler/gen_wasm/README.md
View file

@ -3,42 +3,57 @@
## Plan
- Initial bringup
- Get a wasm backend working for some of the number tests.
- Use a separate `gen_wasm` directory for now, to avoid trying to do bringup and integration at the same time.
- [x] Get a wasm backend working for some of the number tests.
- [x] Use a separate `gen_wasm` directory for now, to avoid trying to do bringup and integration at the same time.
- Get the fundamentals working
- [x] Come up with a way to do control flow
- [x] Flesh out the details of value representations between local variables and stack memory
- [x] Set up a way to write tests with any return value rather than just i64 and f64
- [x] Implement stack memory
- [x] Push and pop stack frames
- [x] Deal with returning structs
- [x] Distinguish which variables go in locals, own stack frame, caller stack frame, etc.
- [ ] Ensure early Return statements don't skip stack cleanup
- [ ] Vendor-in parity_wasm library so that we can use `bumpalo::Vec`
- [ ] Implement relocations
- Requires knowing the _byte_ offset of each call site. This is awkward as the backend builds a `Vec<Instruction>` rather than a `Vec<u8>`. It may be worth serialising each instruction as it is inserted.
- Refactor for code sharing with CPU backends
- [ ] Implement a `scan_ast` pre-pass like `Backend` does, but for reusing Wasm locals rather than CPU registers
- [ ] Extract a trait from `WasmBackend` that looks as similar as possible to `Backend`, to prepare for code sharing
- [ ] Refactor to actually share code between `WasmBackend` and `Backend` if it seems feasible
- Integration
- Move wasm files to `gen_dev/src/wasm`
- Share tests between wasm and x64, with some way of saying which tests work on which backends, and dispatching to different eval helpers based on that.
- Get `build_module` in object_builder.rs to dispatch to the wasm generator (adding some Wasm options to the `Triple` struct)
- Get `build_module` to write to a file, or maybe return `Vec<u8>`, instead of returning an Object structure
- Code sharing
- Try to ensure that both Wasm and x64 use the same `Backend` trait so that we can share code.
- We need to work towards this after we've progressed a bit more with Wasm and gained more understanding and experience of the differences.
- We will have to think about how to deal with the `Backend` code that doesn't apply to Wasm. Perhaps we will end up with more traits like `RegisterBackend` / `StackBackend` or `NativeBackend` / `WasmBackend`, and perhaps even some traits to do with backends that support jumps and those that don't.
## Structured control flow
🚨 **This is an area that could be tricky** 🚨
One of the security features of WebAssembly is that it does not allow unrestricted "jumps" to anywhere you like. It does not have an instruction for that. All of the [control instructions][control-inst] can only implement "structured" control flow, and have names like `if`, `loop`, `block` that you'd normally associate with high-level languages. There are branch (`br`) instructions that can jump to labelled blocks within the same function, but the blocks have to be nested in sensible ways.
[control-inst]: https://webassembly.github.io/spec/core/syntax/instructions.html#control-instructions
Implications:
This way of representing control flow is similar to parts of the Roc AST like `When`, `If` and `LetRec`. But Mono IR converts this to jumps and join points, which are more of a Control Flow Graph than a tree. We need to map back from graph to a tree again in the Wasm backend.
Roc, like most modern languages, is already enforcing structured control flow in the source program. Constructs from the Roc AST like `When`, `If` and `LetRec` can all be converted straightforwardly to Wasm constructs.
Our solution is to wrap all joinpoint/jump graphs in an outer `loop`, with nested `block`s inside it.
However the Mono IR converts this to jumps and join points, which are more of a Control Flow Graph than a tree. That doesn't map so directly to the Wasm structures. This is such a common issue for compiler back-ends that the WebAssembly compiler toolkit `binaryen` has an [API for control-flow graphs][cfg-api]. We're not using `binaryen` right now. It's a C++ library, though it does have a (very thin and somewhat hard-to-use) [Rust wrapper][binaryen-rs]. We should probably investigate this area sooner rather than later. If relooping turns out to be necessary or difficult, we might need to switch from parity_wasm to binaryen.
### Possible future optimisations
> By the way, it's not obvious how to pronounce "binaryen" but apparently it rhymes with "Targaryen", the family name from the "Game of Thrones" TV series
There are other algorithms available that may result in more optimised control flow. We are not focusing on that for our development backend, but here are some notes for future reference.
The WebAssembly compiler toolkit `binaryen` has an [API for control-flow graphs][cfg-api]. We're not using `binaryen` right now. It's a C++ library, though it does have a (very thin and somewhat hard-to-use) [Rust wrapper][binaryen-rs]. Binaryen's control-flow graph API implements the "Relooper" algorithm developed by the Emscripten project and described in [this paper](https://github.com/emscripten-core/emscripten/blob/main/docs/paper.pdf).
> By the way, apparently "binaryen" rhymes with "Targaryen", the family name from the "Game of Thrones" TV series
There is also an improvement on Relooper called ["Stackifier"](https://medium.com/leaningtech/solving-the-structured-control-flow-problem-once-and-for-all-5123117b1ee2). It can reorder the joinpoints and jumps to make code more efficient. (It is also has things Roc wouldn't need but C++ does, like support for "irreducible" graphs that include `goto`).
[cfg-api]: https://github.com/WebAssembly/binaryen/wiki/Compiling-to-WebAssembly-with-Binaryen#cfg-api
[binaryen-rs]: https://crates.io/crates/binaryen
Binaryen's control-flow graph API implements the "Relooper" algorithm developed by the Emscripten project and described in [this paper](https://github.com/emscripten-core/emscripten/blob/main/docs/paper.pdf).
There is an alternative algorithm that is supposed to be an improvement on Relooper, called ["Stackifier"](https://medium.com/leaningtech/solving-the-structured-control-flow-problem-once-and-for-all-5123117b1ee2).
## Stack machine vs register machine
Wasm's instruction set is based on a stack-machine VM. Whereas CPU instructions have named registers that they operate on, Wasm has no named registers at all. The instructions don't contain register names. Instructions can oly operate on whatever data is at the top of the stack.
@ -84,29 +99,30 @@ The Mono IR contains two functions, `Num.add` and `main`, so we generate two cor
(func (;1;) (result i64) ; declare function index 1 (main) with no parameters and an i64 result
(local i64 i64 i64 i64) ; declare 4 local variables, all with type i64, one for each symbol in the Mono IR
i64.const 1 ; load constant of type i64 and value 1 stack=[1]
local.set 0 ; store top of stack to local0 stack=[] local0=1
i64.const 2 ; load constant of type i64 and value 2 stack=[2] local0=1
local.set 1 ; store top of stack to local1 stack=[] local0=1 local1=2
local.get 0 ; load local0 to top of stack stack=[1] local0=1 local1=2
local.get 1 ; load local1 to top of stack stack=[1,2] local0=1 local1=2
call 0 ; call function index 0 (which pops 2 and pushes 1) stack=[3] local0=1 local1=2
local.set 2 ; store top of stack to local2 stack=[] local0=1 local1=2 local2=3
i64.const 4 ; load constant of type i64 and value 4 stack=[4] local0=1 local1=2 local2=3
local.set 3 ; store top of stack to local3 stack=[] local0=1 local1=2 local2=3 local3=4
local.get 2 ; load local2 to top of stack stack=[3] local0=1 local1=2 local2=3 local3=4
local.get 3 ; load local3 to top of stack stack=[3,4] local0=1 local1=2 local2=3 local3=4
call 0 ; call function index 0 (which pops 2 and pushes 1) stack=[7] local0=1 local1=2 local2=3 local3=4
i64.const 1 ; stack=[1]
local.set 0 ; stack=[] local0=1
i64.const 2 ; stack=[2] local0=1
local.set 1 ; stack=[] local0=1 local1=2
local.get 0 ; stack=[1] local0=1 local1=2
local.get 1 ; stack=[1,2] local0=1 local1=2
call 0 ; stack=[3] local0=1 local1=2
local.set 2 ; stack=[] local0=1 local1=2 local2=3
i64.const 4 ; stack=[4] local0=1 local1=2 local2=3
local.set 3 ; stack=[] local0=1 local1=2 local2=3 local3=4
local.get 2 ; stack=[3] local0=1 local1=2 local2=3 local3=4
local.get 3 ; stack=[3,4] local0=1 local1=2 local2=3 local3=4
call 0 ; stack=[7] local0=1 local1=2 local2=3 local3=4
return) ; return the value at the top of the stack
```
If we run this code through the `wasm-opt` tool from the [binaryen toolkit](https://github.com/WebAssembly/binaryen#tools), the unnecessary locals get optimised away. The command line below runs the minimum number of passes to achieve this (`--simplify-locals` must come first).
If we run this code through the `wasm-opt` tool from the [binaryen toolkit](https://github.com/WebAssembly/binaryen#tools), the unnecessary locals get optimised away (which is all of them in this example!). The command line below runs the minimum number of passes to achieve this (`--simplify-locals` must come first).
```
$ wasm-opt --simplify-locals --reorder-locals --vacuum example.wasm > opt.wasm
```
The optimised functions have no local variables, and the code shrinks to about 60% of its original size.
The optimised functions have no local variables at all for this example. (Of course, this is an oversimplified toy example! It might not be so extreme in a real program.)
```
(func (;0;) (param i64 i64) (result i64)
local.get 0
@ -116,9 +132,20 @@ The optimised functions have no local variables, and the code shrinks to about 6
i64.const 1
i64.const 2
call 0
i64.const 4)
i64.const 4
call 0)
```
### Reducing sets and gets
It would be nice to find some cheap optimisation to reduce the number of `local.set` and `local.get` instructions.
We don't need a `local` if the value we want is already at the top of the VM stack. In fact, for our example above, it just so happens that if we simply skip generating the `local.set` instructions, everything _does_ appear on the VM stack in the right order, which means we can skip the `local.get` too. It ends up being very close to the fully optimised version! I assume this is because the Mono IR within the function is in dependency order, but I'm not sure...
Of course the trick is to do this reliably for more complex dependency graphs. I am investigating whether we can do it by optimistically assuming it's OK not to create a local, and then keeping track of which symbols are at which positions in the VM stack after every instruction. Then when we need to use a symbol we can first check if it's on the VM stack and only create a local if it's not. In cases where we _do_ need to create a local, we need to go back and insert a `local.set` instruction at an earlier point in the program. We can make this fast by waiting to do all of the insertions in one batch when we're finalising the procedure.
For a while we thought it would be very helpful to reuse the same local for multiple symbols at different points in the program. And we already have similar code in the CPU backends for register allocation. But on further examination, it doesn't actually buy us much! In our example above, we would still have the same number of `local.set` and `local.get` instructions - they'd just be operating on two locals instead of four! That doesn't shrink much code. Only the declaration at the top of the function would shrink from `(local i64 i64 i64 i64)` to `(local i64 i64)`... and in fact that's only smaller in the text format, it's the same size in the binary format! So the `scan_ast` pass doesn't seem worthwhile for Wasm.
## Memory
WebAssembly programs have a "linear memory" for storing data, which is a block of memory assigned to it by the host. You can assign a min and max size to the memory, and the WebAssembly program can request 64kB pages from the host, just like a "normal" program would request pages from the OS. Addresses start at zero and go up to whatever the current size is. Zero is a perfectly normal address like any other, and dereferencing it is not a segfault. But addresses beyond the current memory size are out of bounds and dereferencing them will cause a panic.
@ -137,7 +164,7 @@ When we are talking about how we store values in _memory_, I'll use the term _st
Of course our program can use another area of memory as a heap as well. WebAssembly doesn't mind how you divide up your memory. It just gives you some memory and some instructions for loading and storing.
## Function calls
## Calling conventions & stack memory
In WebAssembly you call a function by pushing arguments to the stack and then issuing a `call` instruction, which specifies a function index. The VM knows how many values to pop off the stack by examining the _type_ of the function. In our example earlier, `Num.add` had the type `[i64 i64] → [i64]` so it expects to find two i64's on the stack and pushes one i64 back as the result. Remember, the runtime engine will validate the module before running it, and if your generated code is trying to call a function at a point in the program where the wrong value types are on the stack, it will fail validation.
@ -145,11 +172,17 @@ Function arguments are restricted to the four value types, `i32`, `i64`, `f32` a
That's all great for primitive values but what happens when we want to pass more complex data structures between functions?
Well, remember, "stack memory" is not a special kind of memory in WebAssembly, it's just an area of our memory where we _decide_ that we want to implement a stack data structure. So we can implement it however we want. A good choice would be to make our stack frame look the same as it would when we're targeting a CPU, except without the return address (since there's no need for one). We can also decide to pass numbers through the machine stack rather than in stack memory, since that takes fewer instructions.
Well, remember, "stack memory" is not a special kind of memory in WebAssembly, and is separate from the VM stack. It's just an area of our memory where we implement a stack data structure. But there are some conventions that it makes sense to follow so that we can easily link to Wasm code generated from Zig or other languages.
The only other thing we need is a stack pointer. On CPU targets, there's often have a specific "stack pointer" register. WebAssembly has no equivalent to that, but we can use a `global` variable.
### Observations from compiled C code
The system I've outlined above is based on my experience of compiling C to WebAssembly via the Emscripten toolchain (which is built on top of clang). It's also in line with what the WebAssembly project describes [here](https://github.com/WebAssembly/design/blob/main/Rationale.md#locals).
- `global 0` is used as the stack pointer, and its value is normally copied to a `local` as well (presumably because locals tend to be assigned to CPU registers)
- Stack memory grows downwards
- If a C function returns a struct, the compiled WebAssembly function has no return value, but instead has an extra _argument_. The argument is an `i32` pointer to space allocated in the caller's stack, that the called function can write to.
- There is no maximum number of arguments for a WebAssembly function, and arguments are not passed via _stack memory_. This makes sense because the _VM stack_ has no size limit. It's like having a CPU with an unlimited number of registers.
- Stack memory is only used for allocating local variables, not for passing arguments. And it's only used for values that cannot be stored in one of WebAssembly's primitive values (`i32`, `i64`, `f32`, `f64`).
These observations are based on experiments compiling C to WebAssembly via the Emscripten toolchain (which is built on top of clang). It's also in line with what the WebAssembly project describes [here](https://github.com/WebAssembly/design/blob/main/Rationale.md#locals).
## Modules vs Instances

501
compiler/gen_wasm/src/backend.rs
View file

@ -1,5 +1,5 @@
use parity_wasm::builder;
use parity_wasm::builder::{CodeLocation, ModuleBuilder};
use parity_wasm::builder::{CodeLocation, FunctionDefinition, ModuleBuilder, SignatureBuilder};
use parity_wasm::elements::{
BlockType, Instruction, Instruction::*, Instructions, Local, ValueType,
};
@ -10,47 +10,26 @@ use roc_module::symbol::Symbol;
use roc_mono::ir::{CallType, Expr, JoinPointId, Literal, Proc, Stmt};
use roc_mono::layout::{Builtin, Layout};
use crate::layout::WasmLayout;
use crate::storage::SymbolStorage;
use crate::{
copy_memory, pop_stack_frame, push_stack_frame, round_up_to_alignment, LocalId, PTR_SIZE,
PTR_TYPE,
};
// Don't allocate any constant data at address zero or near it. Would be valid, but bug-prone.
// Follow Emscripten's example by using 1kB (4 bytes would probably do)
const UNUSED_DATA_SECTION_BYTES: u32 = 1024;
#[derive(Clone, Copy, Debug)]
struct LocalId(u32);
#[derive(Clone, Copy, Debug)]
struct LabelId(u32);
#[derive(Debug)]
struct SymbolStorage(LocalId, WasmLayout);
#[derive(Clone, Copy, Debug)]
struct WasmLayout {
value_type: ValueType,
stack_memory: u32,
}
impl WasmLayout {
fn new(layout: &Layout) -> Result<Self, String> {
match layout {
Layout::Builtin(Builtin::Int1 | Builtin::Int8 | Builtin::Int16 | Builtin::Int32) => {
Ok(Self {
value_type: ValueType::I32,
stack_memory: 0,
})
}
Layout::Builtin(Builtin::Int64) => Ok(Self {
value_type: ValueType::I64,
stack_memory: 0,
}),
Layout::Builtin(Builtin::Float64) => Ok(Self {
value_type: ValueType::F64,
stack_memory: 0,
}),
x => Err(format!("layout, {:?}, not implemented yet", x)),
}
}
enum LocalKind {
Parameter,
Variable,
}
// TODO: use Bumpalo Vec once parity_wasm supports general iterators (>=0.43)
pub struct WasmBackend<'a> {
// Module: Wasm AST
pub builder: ModuleBuilder,
@ -62,12 +41,12 @@ pub struct WasmBackend<'a> {
// Functions: Wasm AST
instructions: std::vec::Vec<Instruction>,
ret_type: ValueType,
arg_types: std::vec::Vec<ValueType>,
locals: std::vec::Vec<Local>,
// Functions: internal state & IR mappings
stack_memory: u32,
stack_memory: i32,
stack_frame_pointer: Option<LocalId>,
symbol_storage_map: MutMap<Symbol, SymbolStorage>,
/// how many blocks deep are we (used for jumps)
block_depth: u32,
@ -87,12 +66,12 @@ impl<'a> WasmBackend<'a> {
// Functions: Wasm AST
instructions: std::vec::Vec::with_capacity(256),
ret_type: ValueType::I32,
arg_types: std::vec::Vec::with_capacity(8),
locals: std::vec::Vec::with_capacity(32),
// Functions: internal state & IR mappings
stack_memory: 0,
stack_frame_pointer: None,
symbol_storage_map: MutMap::default(),
block_depth: 0,
joinpoint_label_map: MutMap::default(),
@ -107,48 +86,18 @@ impl<'a> WasmBackend<'a> {
// Functions: internal state & IR mappings
self.stack_memory = 0;
self.stack_frame_pointer = None;
self.symbol_storage_map.clear();
// joinpoint_label_map.clear();
self.joinpoint_label_map.clear();
assert_eq!(self.block_depth, 0);
}
pub fn build_proc(&mut self, proc: Proc<'a>, sym: Symbol) -> Result<u32, String> {
let ret_layout = WasmLayout::new(&proc.ret_layout)?;
if ret_layout.stack_memory > 0 {
// TODO: if returning a struct by value, add an extra argument for a pointer to callee's stack memory
return Err(format!(
"Not yet implemented: Return in stack memory for non-primtitive layouts like {:?}",
proc.ret_layout
));
}
self.ret_type = ret_layout.value_type;
self.arg_types.reserve(proc.args.len());
for (layout, symbol) in proc.args {
let wasm_layout = WasmLayout::new(layout)?;
self.arg_types.push(wasm_layout.value_type);
self.insert_local(wasm_layout, *symbol);
}
let signature_builder = self.start_proc(&proc);
self.build_stmt(&proc.body, &proc.ret_layout)?;
let signature = builder::signature()
.with_params(self.arg_types.clone()) // requires std::Vec, not Bumpalo
.with_result(self.ret_type)
.build_sig();
// functions must end with an End instruction/opcode
let mut instructions = self.instructions.clone();
instructions.push(Instruction::End);
let function_def = builder::function()
.with_signature(signature)
.body()
.with_locals(self.locals.clone())
.with_instructions(Instructions::new(instructions))
.build() // body
.build(); // function
let function_def = self.finalize_proc(signature_builder);
let location = self.builder.push_function(function_def);
let function_index = location.body;
self.proc_symbol_map.insert(sym, location);
@ -157,32 +106,169 @@ impl<'a> WasmBackend<'a> {
Ok(function_index)
}
fn insert_local(&mut self, layout: WasmLayout, symbol: Symbol) -> LocalId {
self.stack_memory += layout.stack_memory;
let index = self.symbol_storage_map.len();
if index >= self.arg_types.len() {
self.locals.push(Local::new(1, layout.value_type));
fn start_proc(&mut self, proc: &Proc<'a>) -> SignatureBuilder {
let ret_layout = WasmLayout::new(&proc.ret_layout);
let signature_builder = if let WasmLayout::StackMemory { .. } = ret_layout {
self.arg_types.push(PTR_TYPE);
self.start_block(BlockType::NoResult); // block to ensure all paths pop stack memory (if any)
builder::signature()
} else {
let ret_type = ret_layout.value_type();
self.start_block(BlockType::Value(ret_type)); // block to ensure all paths pop stack memory (if any)
builder::signature().with_result(ret_type)
};
for (layout, symbol) in proc.args {
self.insert_local(WasmLayout::new(layout), *symbol, LocalKind::Parameter);
}
let local_id = LocalId(index as u32);
let storage = SymbolStorage(local_id, layout);
self.symbol_storage_map.insert(symbol, storage);
local_id
signature_builder.with_params(self.arg_types.clone())
}
fn get_symbol_storage(&self, sym: &Symbol) -> Result<&SymbolStorage, String> {
self.symbol_storage_map.get(sym).ok_or_else(|| {
format!(
fn finalize_proc(&mut self, signature_builder: SignatureBuilder) -> FunctionDefinition {
self.end_block(); // end the block from start_proc, to ensure all paths pop stack memory (if any)
let mut final_instructions = Vec::with_capacity(self.instructions.len() + 10);
if self.stack_memory > 0 {
push_stack_frame(
&mut final_instructions,
self.stack_memory,
self.stack_frame_pointer.unwrap(),
);
}
final_instructions.extend(self.instructions.drain(0..));
if self.stack_memory > 0 {
pop_stack_frame(
&mut final_instructions,
self.stack_memory,
self.stack_frame_pointer.unwrap(),
);
}
final_instructions.push(End);
builder::function()
.with_signature(signature_builder.build_sig())
.body()
.with_locals(self.locals.clone())
.with_instructions(Instructions::new(final_instructions))
.build() // body
.build() // function
}
fn insert_local(
&mut self,
wasm_layout: WasmLayout,
symbol: Symbol,
kind: LocalKind,
) -> SymbolStorage {
let local_id = LocalId((self.arg_types.len() + self.locals.len()) as u32);
let storage = match kind {
LocalKind::Parameter => {
// Already stack-allocated by the caller if needed.
self.arg_types.push(wasm_layout.value_type());
match wasm_layout {
WasmLayout::LocalOnly(value_type, size) => SymbolStorage::ParamPrimitive {
local_id,
value_type,
size,
},
WasmLayout::HeapMemory => SymbolStorage::ParamPrimitive {
local_id,
value_type: PTR_TYPE,
size: PTR_SIZE,
},
WasmLayout::StackMemory {
size,
alignment_bytes,
} => SymbolStorage::ParamStackMemory {
local_id,
size,
alignment_bytes,
},
}
}
LocalKind::Variable => {
self.locals.push(Local::new(1, wasm_layout.value_type()));
match wasm_layout {
WasmLayout::LocalOnly(value_type, size) => SymbolStorage::VarPrimitive {
local_id,
value_type,
size,
},
WasmLayout::HeapMemory => SymbolStorage::VarHeapMemory { local_id },
WasmLayout::StackMemory {
size,
alignment_bytes,
} => {
let offset =
round_up_to_alignment(self.stack_memory, alignment_bytes as i32);
self.stack_memory = offset + size as i32;
match self.stack_frame_pointer {
None => {
// This is the first stack-memory variable in the function
// That means we can reuse it as the stack frame pointer,
// and it will get initialised at the start of the function
self.stack_frame_pointer = Some(local_id);
}
Some(frame_ptr_id) => {
// This local points to the base of a struct, at an offset from the stack frame pointer
// Having one local per variable means params and locals work the same way in code gen.
// (alternatively we could use one frame pointer + offset for all struct variables)
self.instructions.extend([
GetLocal(frame_ptr_id.0),
I32Const(offset),
I32Add,
SetLocal(local_id.0),
]);
}
};
SymbolStorage::VarStackMemory {
local_id,
size,
offset: offset as u32,
alignment_bytes,
}
}
}
}
};
self.symbol_storage_map.insert(symbol, storage.clone());
storage
}
fn get_symbol_storage(&self, sym: &Symbol) -> &SymbolStorage {
self.symbol_storage_map.get(sym).unwrap_or_else(|| {
panic!(
"Symbol {:?} not found in function scope:\n{:?}",
sym, self.symbol_storage_map
)
})
}
fn load_from_symbol(&mut self, sym: &Symbol) -> Result<(), String> {
let SymbolStorage(LocalId(local_id), _) = self.get_symbol_storage(sym)?;
let id: u32 = *local_id;
self.instructions.push(GetLocal(id));
Ok(())
fn local_id_from_symbol(&self, sym: &Symbol) -> LocalId {
let storage = self.get_symbol_storage(sym);
storage.local_id()
}
fn load_symbol(&mut self, sym: &Symbol) {
let storage = self.get_symbol_storage(sym);
let index: u32 = storage.local_id().0;
self.instructions.push(GetLocal(index));
}
/// start a loop that leaves a value on the stack
@ -193,12 +279,9 @@ impl<'a> WasmBackend<'a> {
self.instructions.push(Loop(BlockType::Value(value_type)));
}
fn start_block(&mut self) {
fn start_block(&mut self, block_type: BlockType) {
self.block_depth += 1;
// Our blocks always end with a `return` or `br`,
// so they never leave extra values on the stack
self.instructions.push(Block(BlockType::NoResult));
self.instructions.push(Block(block_type));
}
fn end_block(&mut self) {
@ -208,36 +291,77 @@ impl<'a> WasmBackend<'a> {
fn build_stmt(&mut self, stmt: &Stmt<'a>, ret_layout: &Layout<'a>) -> Result<(), String> {
match stmt {
// This pattern is a simple optimisation to get rid of one local and two instructions per proc.
// If we are just returning the expression result, then don't SetLocal and immediately GetLocal
// Simple optimisation: if we are just returning the expression, we don't need a local
Stmt::Let(let_sym, expr, layout, Stmt::Ret(ret_sym)) if let_sym == ret_sym => {
let wasm_layout = WasmLayout::new(layout);
if let WasmLayout::StackMemory {
size,
alignment_bytes,
} = wasm_layout
{
// Map this symbol to the first argument (pointer into caller's stack)
// Saves us from having to copy it later
let storage = SymbolStorage::ParamStackMemory {
local_id: LocalId(0),
size,
alignment_bytes,
};
self.symbol_storage_map.insert(*let_sym, storage);
}
self.build_expr(let_sym, expr, layout)?;
self.instructions.push(Return);
self.instructions.push(Br(self.block_depth)); // jump to end of function (stack frame pop)
Ok(())
}
Stmt::Let(sym, expr, layout, following) => {
let wasm_layout = WasmLayout::new(layout)?;
let local_id = self.insert_local(wasm_layout, *sym);
let wasm_layout = WasmLayout::new(layout);
let local_id = self
.insert_local(wasm_layout, *sym, LocalKind::Variable)
.local_id();
self.build_expr(sym, expr, layout)?;
self.instructions.push(SetLocal(local_id.0));
// If this local is shared with the stack frame pointer, it's already assigned
match self.stack_frame_pointer {
Some(sfp) if sfp == local_id => {}
_ => self.instructions.push(SetLocal(local_id.0)),
}
self.build_stmt(following, ret_layout)?;
Ok(())
}
Stmt::Ret(sym) => {
if let Some(SymbolStorage(local_id, _)) = self.symbol_storage_map.get(sym) {
self.instructions.push(GetLocal(local_id.0));
self.instructions.push(Return);
Ok(())
} else {
Err(format!(
"Not yet implemented: returning values with layout {:?}",
ret_layout
))
use crate::storage::SymbolStorage::*;
let storage = self.symbol_storage_map.get(sym).unwrap();
match storage {
VarStackMemory {
local_id,
size,
alignment_bytes,
..
}
| ParamStackMemory {
local_id,
size,
alignment_bytes,
} => {
let from = *local_id;
let to = LocalId(0);
copy_memory(&mut self.instructions, from, to, *size, *alignment_bytes, 0);
}
ParamPrimitive { local_id, .. }
| VarPrimitive { local_id, .. }
| VarHeapMemory { local_id, .. } => {
self.instructions.push(GetLocal(local_id.0));
self.instructions.push(Br(self.block_depth)); // jump to end of function (for stack frame pop)
}
}
Ok(())
}
Stmt::Switch {
@ -253,19 +377,16 @@ impl<'a> WasmBackend<'a> {
// create (number_of_branches - 1) new blocks.
for _ in 0..branches.len() {
self.start_block()
self.start_block(BlockType::NoResult)
}
// the LocalId of the symbol that we match on
let matched_on = match self.symbol_storage_map.get(cond_symbol) {
Some(SymbolStorage(local_id, _)) => local_id.0,
None => unreachable!("symbol not defined: {:?}", cond_symbol),
};
let matched_on = self.local_id_from_symbol(cond_symbol);
// then, we jump whenever the value under scrutiny is equal to the value of a branch
for (i, (value, _, _)) in branches.iter().enumerate() {
// put the cond_symbol on the top of the stack
self.instructions.push(GetLocal(matched_on));
self.instructions.push(GetLocal(matched_on.0));
self.instructions.push(I32Const(*value as i32));
@ -299,13 +420,15 @@ impl<'a> WasmBackend<'a> {
// make locals for join pointer parameters
let mut jp_parameter_local_ids = std::vec::Vec::with_capacity(parameters.len());
for parameter in parameters.iter() {
let wasm_layout = WasmLayout::new(&parameter.layout)?;
let local_id = self.insert_local(wasm_layout, parameter.symbol);
let wasm_layout = WasmLayout::new(&parameter.layout);
let local_id = self
.insert_local(wasm_layout, parameter.symbol, LocalKind::Variable)
.local_id();
jp_parameter_local_ids.push(local_id);
}
self.start_block();
self.start_block(BlockType::NoResult);
self.joinpoint_label_map
.insert(*id, (self.block_depth, jp_parameter_local_ids));
@ -316,8 +439,8 @@ impl<'a> WasmBackend<'a> {
// A `return` inside of a `loop` seems to make it so that the `loop` itself
// also "returns" (so, leaves on the stack) a value of the return type.
let return_wasm_layout = WasmLayout::new(ret_layout)?;
self.start_loop_with_return(return_wasm_layout.value_type);
let return_wasm_layout = WasmLayout::new(ret_layout);
self.start_loop_with_return(return_wasm_layout.value_type());
self.build_stmt(body, ret_layout)?;
@ -331,12 +454,8 @@ impl<'a> WasmBackend<'a> {
// put the arguments on the stack
for (symbol, local_id) in arguments.iter().zip(locals.iter()) {
let argument = match self.symbol_storage_map.get(symbol) {
Some(SymbolStorage(local_id, _)) => local_id.0,
None => unreachable!("symbol not defined: {:?}", symbol),
};
self.instructions.push(GetLocal(argument));
let argument = self.local_id_from_symbol(symbol);
self.instructions.push(GetLocal(argument.0));
self.instructions.push(SetLocal(local_id.0));
}
@ -365,7 +484,7 @@ impl<'a> WasmBackend<'a> {
}) => match call_type {
CallType::ByName { name: func_sym, .. } => {
for arg in *arguments {
self.load_from_symbol(arg)?;
self.load_symbol(arg);
}
let function_location = self.proc_symbol_map.get(func_sym).ok_or(format!(
"Cannot find function {:?} called from {:?}",
@ -381,38 +500,112 @@ impl<'a> WasmBackend<'a> {
x => Err(format!("the call type, {:?}, is not yet implemented", x)),
},
Expr::Struct(fields) => self.create_struct(sym, layout, fields),
x => Err(format!("Expression is not yet implemented {:?}", x)),
}
}
fn load_literal(&mut self, lit: &Literal<'a>, layout: &Layout<'a>) -> Result<(), String> {
match lit {
Literal::Bool(x) => {
self.instructions.push(I32Const(*x as i32));
Ok(())
}
Literal::Byte(x) => {
self.instructions.push(I32Const(*x as i32));
Ok(())
}
Literal::Int(x) => {
match layout {
Layout::Builtin(Builtin::Int32) => {
self.instructions.push(I32Const(*x as i32));
}
Layout::Builtin(Builtin::Int64) => {
self.instructions.push(I64Const(*x as i64));
}
x => panic!("loading literal, {:?}, is not yet implemented", x),
let instruction = match lit {
Literal::Bool(x) => I32Const(*x as i32),
Literal::Byte(x) => I32Const(*x as i32),
Literal::Int(x) => match layout {
Layout::Builtin(Builtin::Int64) => I64Const(*x as i64),
Layout::Builtin(
Builtin::Int32
| Builtin::Int16
| Builtin::Int8
| Builtin::Int1
| Builtin::Usize,
) => I32Const(*x as i32),
x => {
return Err(format!("loading literal, {:?}, is not yet implemented", x));
}
Ok(())
},
Literal::Float(x) => match layout {
Layout::Builtin(Builtin::Float64) => F64Const((*x as f64).to_bits()),
Layout::Builtin(Builtin::Float32) => F32Const((*x as f32).to_bits()),
x => {
return Err(format!("loading literal, {:?}, is not yet implemented", x));
}
},
x => {
return Err(format!("loading literal, {:?}, is not yet implemented", x));
}
Literal::Float(x) => {
let val: f64 = *x;
self.instructions.push(F64Const(val.to_bits()));
Ok(())
};
self.instructions.push(instruction);
Ok(())
}
fn create_struct(
&mut self,
sym: &Symbol,
layout: &Layout<'a>,
fields: &'a [Symbol],
) -> Result<(), String> {
let storage = self.get_symbol_storage(sym).to_owned();
if let Layout::Struct(field_layouts) = layout {
match storage {
SymbolStorage::VarStackMemory { local_id, size, .. }
| SymbolStorage::ParamStackMemory { local_id, size, .. } => {
if size > 0 {
let mut relative_offset = 0;
for (field, _) in fields.iter().zip(field_layouts.iter()) {
relative_offset += self.copy_symbol_to_pointer_at_offset(
local_id,
relative_offset,
field,
);
}
} else {
return Err(format!("Not supported yet: zero-size struct at {:?}", sym));
}
}
_ => {
return Err(format!(
"Cannot create struct {:?} with storage {:?}",
sym, storage
));
}
}
x => Err(format!("loading literal, {:?}, is not yet implemented", x)),
} else {
// Struct expression but not Struct layout => single element. Copy it.
let field_storage = self.get_symbol_storage(&fields[0]).to_owned();
self.copy_storage(&storage, &field_storage);
}
Ok(())
}
fn copy_symbol_to_pointer_at_offset(
&mut self,
to_ptr: LocalId,
to_offset: u32,
from_symbol: &Symbol,
) -> u32 {
let from_storage = self.get_symbol_storage(from_symbol).to_owned();
from_storage.copy_to_memory(&mut self.instructions, to_ptr, to_offset)
}
fn copy_storage(&mut self, to: &SymbolStorage, from: &SymbolStorage) {
let has_stack_memory = to.has_stack_memory();
debug_assert!(from.has_stack_memory() == has_stack_memory);
if !has_stack_memory {
debug_assert!(from.value_type() == to.value_type());
self.instructions.push(GetLocal(from.local_id().0));
self.instructions.push(SetLocal(to.local_id().0));
} else {
let (size, alignment_bytes) = from.stack_size_and_alignment();
copy_memory(
&mut self.instructions,
from.local_id(),
to.local_id(),
size,
alignment_bytes,
0,
);
}
}
@ -423,10 +616,10 @@ impl<'a> WasmBackend<'a> {
return_layout: &Layout<'a>,
) -> Result<(), String> {
for arg in args {
self.load_from_symbol(arg)?;
self.load_symbol(arg);
}
let wasm_layout = WasmLayout::new(return_layout)?;
self.build_instructions_lowlevel(lowlevel, wasm_layout.value_type)?;
let wasm_layout = WasmLayout::new(return_layout);
self.build_instructions_lowlevel(lowlevel, wasm_layout.value_type())?;
Ok(())
}
@ -441,7 +634,7 @@ impl<'a> WasmBackend<'a> {
// For those, we'll need to pre-process each argument before the main op,
// so simple arrays of instructions won't work. But there are common patterns.
let instructions: &[Instruction] = match lowlevel {
// Wasm type might not be enough, may need to sign-extend i8 etc. Maybe in load_from_symbol?
// Wasm type might not be enough, may need to sign-extend i8 etc. Maybe in load_symbol?
LowLevel::NumAdd => match return_value_type {
ValueType::I32 => &[I32Add],
ValueType::I64 => &[I64Add],

82
compiler/gen_wasm/src/layout.rs Normal file
View file

@ -0,0 +1,82 @@
use parity_wasm::elements::ValueType;
use roc_mono::layout::{Layout, UnionLayout};
use crate::{PTR_SIZE, PTR_TYPE};
// See README for background information on Wasm locals, memory and function calls
#[derive(Debug, Clone)]
pub enum WasmLayout {
// Primitive number value. Just a Wasm local, without any stack memory.
// For example, Roc i8 is represented as Wasm i32. Store the type and the original size.
LocalOnly(ValueType, u32),
// Local pointer to stack memory
StackMemory { size: u32, alignment_bytes: u32 },
// Local pointer to heap memory
HeapMemory,
}
impl WasmLayout {
pub fn new(layout: &Layout) -> Self {
use roc_mono::layout::Builtin::*;
use UnionLayout::*;
use ValueType::*;
let size = layout.stack_size(PTR_SIZE);
let alignment_bytes = layout.alignment_bytes(PTR_SIZE);
match layout {
Layout::Builtin(Int32 | Int16 | Int8 | Int1 | Usize) => Self::LocalOnly(I32, size),
Layout::Builtin(Int64) => Self::LocalOnly(I64, size),
Layout::Builtin(Float32) => Self::LocalOnly(F32, size),
Layout::Builtin(Float64) => Self::LocalOnly(F64, size),
Layout::Builtin(
Int128
| Decimal
| Float128
| Str
| Dict(_, _)
| Set(_)
| List(_)
| EmptyStr
| EmptyList
| EmptyDict
| EmptySet,
)
| Layout::Struct(_)
| Layout::LambdaSet(_)
| Layout::Union(NonRecursive(_)) => Self::StackMemory {
size,
alignment_bytes,
},
Layout::Union(
Recursive(_)
| NonNullableUnwrapped(_)
| NullableWrapped { .. }
| NullableUnwrapped { .. },
)
| Layout::RecursivePointer => Self::HeapMemory,
}
}
pub fn value_type(&self) -> ValueType {
match self {
Self::LocalOnly(type_, _) => *type_,
_ => PTR_TYPE,
}
}
#[allow(dead_code)]
pub fn stack_memory(&self) -> u32 {
match self {
Self::StackMemory { size, .. } => *size,
_ => 0,
}
}
}

133
compiler/gen_wasm/src/lib.rs
View file

@ -1,9 +1,11 @@
mod backend;
pub mod from_wasm32_memory;
mod layout;
mod storage;
use bumpalo::Bump;
use parity_wasm::builder;
use parity_wasm::elements::Internal;
use parity_wasm::elements::{Instruction, Instruction::*, Internal, ValueType};
use roc_collections::all::{MutMap, MutSet};
use roc_module::symbol::{Interns, Symbol};
@ -12,6 +14,21 @@ use roc_mono::layout::LayoutIds;
use crate::backend::WasmBackend;
const PTR_SIZE: u32 = 4;
const PTR_TYPE: ValueType = ValueType::I32;
// All usages of these alignment constants take u32, so an enum wouldn't add any safety.
pub const ALIGN_1: u32 = 0;
pub const ALIGN_2: u32 = 1;
pub const ALIGN_4: u32 = 2;
pub const ALIGN_8: u32 = 3;
pub const STACK_POINTER_GLOBAL_ID: u32 = 0;
pub const STACK_ALIGNMENT_BYTES: i32 = 16;
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub struct LocalId(pub u32);
pub struct Env<'a> {
pub arena: &'a Bump, // not really using this much, parity_wasm works with std::vec a lot
pub interns: Interns,
@ -21,7 +38,18 @@ pub struct Env<'a> {
pub fn build_module<'a>(
env: &'a Env,
procedures: MutMap<(Symbol, ProcLayout<'a>), Proc<'a>>,
) -> Result<std::vec::Vec<u8>, String> {
) -> Result<Vec<u8>, String> {
let (builder, _) = build_module_help(env, procedures)?;
let module = builder.build();
module
.to_bytes()
.map_err(|e| -> String { format!("Error serialising Wasm module {:?}", e) })
}
pub fn build_module_help<'a>(
env: &'a Env,
procedures: MutMap<(Symbol, ProcLayout<'a>), Proc<'a>>,
) -> Result<(builder::ModuleBuilder, u32), String> {
let mut backend = WasmBackend::new();
let mut layout_ids = LayoutIds::default();
@ -38,8 +66,9 @@ pub fn build_module<'a>(
let mut procedures: std::vec::Vec<_> = procedures.into_iter().collect();
procedures.sort_by(|a, b| b.0 .0.cmp(&a.0 .0));
let mut function_index: u32 = 0;
for ((sym, layout), proc) in procedures {
let function_index = backend.build_proc(proc, sym)?;
function_index = backend.build_proc(proc, sym)?;
if env.exposed_to_host.contains(&sym) {
let fn_name = layout_ids
.get_toplevel(sym, &layout)
@ -54,6 +83,11 @@ pub fn build_module<'a>(
}
}
// Because of the sorting above, we know the last function in the `for` is the main function.
// Here we grab its index and return it, so that the test_wrapper is able to call it.
// This is a workaround until we implement object files with symbols and relocations.
let main_function_index = function_index;
const MIN_MEMORY_SIZE_KB: u32 = 1024;
const PAGE_SIZE_KB: u32 = 64;
@ -61,15 +95,98 @@ pub fn build_module<'a>(
.with_min(MIN_MEMORY_SIZE_KB / PAGE_SIZE_KB)
.build();
backend.builder.push_memory(memory);
let memory_export = builder::export()
.field("memory")
.with_internal(Internal::Memory(0))
.build();
backend.builder.push_export(memory_export);
let module = backend.builder.build();
module
.to_bytes()
.map_err(|e| -> String { format!("Error serialising Wasm module {:?}", e) })
let stack_pointer_global = builder::global()
.with_type(PTR_TYPE)
.mutable()
.init_expr(Instruction::I32Const((MIN_MEMORY_SIZE_KB * 1024) as i32))
.build();
backend.builder.push_global(stack_pointer_global);
Ok((backend.builder, main_function_index))
}
fn encode_alignment(bytes: u32) -> u32 {
match bytes {
1 => ALIGN_1,
2 => ALIGN_2,
4 => ALIGN_4,
8 => ALIGN_8,
_ => panic!("{:?}-byte alignment is not supported", bytes),
}
}
fn copy_memory(
instructions: &mut Vec<Instruction>,
from_ptr: LocalId,
to_ptr: LocalId,
size: u32,
alignment_bytes: u32,
offset: u32,
) {
let alignment_flag = encode_alignment(alignment_bytes);
let mut current_offset = offset;
while size - current_offset >= 8 {
instructions.push(GetLocal(to_ptr.0));
instructions.push(GetLocal(from_ptr.0));
instructions.push(I64Load(alignment_flag, current_offset));
instructions.push(I64Store(alignment_flag, current_offset));
current_offset += 8;
}
if size - current_offset >= 4 {
instructions.push(GetLocal(to_ptr.0));
instructions.push(GetLocal(from_ptr.0));
instructions.push(I32Load(alignment_flag, current_offset));
instructions.push(I32Store(alignment_flag, current_offset));
current_offset += 4;
}
while size - current_offset > 0 {
instructions.push(GetLocal(to_ptr.0));
instructions.push(GetLocal(from_ptr.0));
instructions.push(I32Load8U(alignment_flag, current_offset));
instructions.push(I32Store8(alignment_flag, current_offset));
current_offset += 1;
}
}
/// Round up to alignment_bytes (assumed to be a power of 2)
pub fn round_up_to_alignment(unaligned: i32, alignment_bytes: i32) -> i32 {
let mut aligned = unaligned;
aligned += alignment_bytes - 1; // if lower bits are non-zero, push it over the next boundary
aligned &= -alignment_bytes; // mask with a flag that has upper bits 1, lower bits 0
aligned
}
pub fn push_stack_frame(
instructions: &mut Vec<Instruction>,
size: i32,
local_frame_pointer: LocalId,
) {
let aligned_size = round_up_to_alignment(size, STACK_ALIGNMENT_BYTES);
instructions.extend([
GetGlobal(STACK_POINTER_GLOBAL_ID),
I32Const(aligned_size),
I32Sub,
TeeLocal(local_frame_pointer.0),
SetGlobal(STACK_POINTER_GLOBAL_ID),
]);
}
pub fn pop_stack_frame(
instructions: &mut Vec<Instruction>,
size: i32,
local_frame_pointer: LocalId,
) {
let aligned_size = round_up_to_alignment(size, STACK_ALIGNMENT_BYTES);
instructions.extend([
GetLocal(local_frame_pointer.0),
I32Const(aligned_size),
I32Add,
SetGlobal(STACK_POINTER_GLOBAL_ID),
]);
}

146
compiler/gen_wasm/src/storage.rs Normal file
View file

@ -0,0 +1,146 @@
use crate::{copy_memory, LocalId, ALIGN_1, ALIGN_2, ALIGN_4, ALIGN_8};
use parity_wasm::elements::{Instruction, Instruction::*, ValueType};
#[derive(Debug, Clone)]
pub enum SymbolStorage {
VarPrimitive {
local_id: LocalId,
value_type: ValueType,
size: u32,
},
ParamPrimitive {
local_id: LocalId,
value_type: ValueType,
size: u32,
},
VarStackMemory {
local_id: LocalId,
size: u32,
offset: u32,
alignment_bytes: u32,
},
ParamStackMemory {
local_id: LocalId,
size: u32,
alignment_bytes: u32,
},
VarHeapMemory {
local_id: LocalId,
},
}
impl SymbolStorage {
pub fn local_id(&self) -> LocalId {
match self {
Self::ParamPrimitive { local_id, .. } => *local_id,
Self::ParamStackMemory { local_id, .. } => *local_id,
Self::VarPrimitive { local_id, .. } => *local_id,
Self::VarStackMemory { local_id, .. } => *local_id,
Self::VarHeapMemory { local_id, .. } => *local_id,
}
}
pub fn value_type(&self) -> ValueType {
match self {
Self::ParamPrimitive { value_type, .. } => *value_type,
Self::VarPrimitive { value_type, .. } => *value_type,
Self::ParamStackMemory { .. } => ValueType::I32,
Self::VarStackMemory { .. } => ValueType::I32,
Self::VarHeapMemory { .. } => ValueType::I32,
}
}
pub fn has_stack_memory(&self) -> bool {
match self {
Self::ParamStackMemory { .. } => true,
Self::VarStackMemory { .. } => true,
Self::ParamPrimitive { .. } => false,
Self::VarPrimitive { .. } => false,
Self::VarHeapMemory { .. } => false,
}
}
pub fn stack_size_and_alignment(&self) -> (u32, u32) {
match self {
Self::VarStackMemory {
size,
alignment_bytes,
..
}
| Self::ParamStackMemory {
size,
alignment_bytes,
..
} => (*size, *alignment_bytes),
_ => (0, 0),
}
}
pub fn copy_to_memory(
&self,
instructions: &mut Vec<Instruction>,
to_pointer: LocalId,
to_offset: u32,
) -> u32 {
match self {
Self::ParamPrimitive {
local_id,
value_type,
size,
..
}
| Self::VarPrimitive {
local_id,
value_type,
size,
..
} => {
let store_instruction = match (value_type, size) {
(ValueType::I64, 8) => I64Store(ALIGN_8, to_offset),
(ValueType::I32, 4) => I32Store(ALIGN_4, to_offset),
(ValueType::I32, 2) => I32Store16(ALIGN_2, to_offset),
(ValueType::I32, 1) => I32Store8(ALIGN_1, to_offset),
(ValueType::F32, 4) => F32Store(ALIGN_4, to_offset),
(ValueType::F64, 8) => F64Store(ALIGN_8, to_offset),
_ => {
panic!("Cannot store {:?} with alignment of {:?}", value_type, size);
}
};
instructions.push(GetLocal(to_pointer.0));
instructions.push(GetLocal(local_id.0));
instructions.push(store_instruction);
*size
}
Self::ParamStackMemory {
local_id,
size,
alignment_bytes,
}
| Self::VarStackMemory {
local_id,
size,
alignment_bytes,
..
} => {
copy_memory(
instructions,
*local_id,
to_pointer,
*size,
*alignment_bytes,
to_offset,
);
*size
}
Self::VarHeapMemory { local_id, .. } => {
instructions.push(GetLocal(to_pointer.0));
instructions.push(GetLocal(local_id.0));
instructions.push(I32Store(ALIGN_4, to_offset));
4
}
}
}
}

52
compiler/gen_wasm/tests/helpers/eval.rs
View file

@ -1,6 +1,10 @@
use roc_can::builtins::builtin_defs_map;
use roc_collections::all::{MutMap, MutSet};
// use roc_std::{RocDec, RocList, RocOrder, RocStr};
use crate::helpers::wasm32_test_result::Wasm32TestResult;
use roc_gen_wasm::from_wasm32_memory::FromWasm32Memory;
const TEST_WRAPPER_NAME: &str = "test_wrapper";
fn promote_expr_to_module(src: &str) -> String {
let mut buffer = String::from("app \"test\" provides [ main ] to \"./platform\"\n\nmain =\n");
@ -16,12 +20,11 @@ fn promote_expr_to_module(src: &str) -> String {
}
#[allow(dead_code)]
pub fn helper_wasm<'a>(
pub fn helper_wasm<'a, T: Wasm32TestResult>(
arena: &'a bumpalo::Bump,
src: &str,
stdlib: &'a roc_builtins::std::StdLib,
_is_gen_test: bool,
_ignore_problems: bool,
_result_type_dummy: &T,
) -> wasmer::Instance {
use std::path::{Path, PathBuf};
@ -91,7 +94,11 @@ pub fn helper_wasm<'a>(
exposed_to_host,
};
let module_bytes = roc_gen_wasm::build_module(&env, procedures).unwrap();
let (mut builder, main_function_index) =
roc_gen_wasm::build_module_help(&env, procedures).unwrap();
T::insert_test_wrapper(&mut builder, TEST_WRAPPER_NAME, main_function_index);
let module_bytes = builder.build().to_bytes().unwrap();
// for debugging (e.g. with wasm2wat)
if false {
@ -128,45 +135,40 @@ pub fn helper_wasm<'a>(
}
#[allow(dead_code)]
pub fn assert_wasm_evals_to_help<T>(src: &str, ignore_problems: bool) -> Result<T, String>
pub fn assert_wasm_evals_to_help<T>(src: &str, expected: T) -> Result<T, String>
where
T: Copy,
T: FromWasm32Memory + Wasm32TestResult,
{
let arena = bumpalo::Bump::new();
// NOTE the stdlib must be in the arena; just taking a reference will segfault
let stdlib = arena.alloc(roc_builtins::std::standard_stdlib());
let is_gen_test = true;
let instance =
crate::helpers::eval::helper_wasm(&arena, src, stdlib, is_gen_test, ignore_problems);
let instance = crate::helpers::eval::helper_wasm(&arena, src, stdlib, &expected);
let main_function = instance.exports.get_function("#UserApp_main_1").unwrap();
let memory = instance.exports.get_memory("memory").unwrap();
match main_function.call(&[]) {
let test_wrapper = instance.exports.get_function(TEST_WRAPPER_NAME).unwrap();
match test_wrapper.call(&[]) {
Err(e) => Err(format!("{:?}", e)),
Ok(result) => {
let integer = match result[0] {
wasmer::Value::I32(a) => a as i64,
wasmer::Value::I64(a) => a,
wasmer::Value::F64(a) => a.to_bits() as i64,
let address = match result[0] {
wasmer::Value::I32(a) => a,
_ => panic!(),
};
let output_ptr: &T;
unsafe {
output_ptr = std::mem::transmute::<&i64, &T>(&integer);
}
let output = <T as FromWasm32Memory>::decode(memory, address as u32);
Ok(*output_ptr)
Ok(output)
}
}
}
#[macro_export]
macro_rules! assert_wasm_evals_to {
($src:expr, $expected:expr, $ty:ty, $transform:expr, $ignore_problems:expr) => {
match $crate::helpers::eval::assert_wasm_evals_to_help::<$ty>($src, $ignore_problems) {
($src:expr, $expected:expr, $ty:ty, $transform:expr) => {
match $crate::helpers::eval::assert_wasm_evals_to_help::<$ty>($src, $expected) {
Err(msg) => println!("{:?}", msg),
Ok(actual) => {
#[allow(clippy::bool_assert_comparison)]
@ -176,11 +178,11 @@ macro_rules! assert_wasm_evals_to {
};
($src:expr, $expected:expr, $ty:ty) => {
$crate::assert_wasm_evals_to!($src, $expected, $ty, $crate::helpers::eval::identity, false);
$crate::assert_wasm_evals_to!($src, $expected, $ty, $crate::helpers::eval::identity);
};
($src:expr, $expected:expr, $ty:ty, $transform:expr) => {
$crate::assert_wasm_evals_to!($src, $expected, $ty, $transform, false);
$crate::assert_wasm_evals_to!($src, $expected, $ty, $transform);
};
}
@ -192,7 +194,7 @@ macro_rules! assert_evals_to {
($src:expr, $expected:expr, $ty:ty, $transform:expr) => {
// Same as above, except with an additional transformation argument.
{
$crate::assert_wasm_evals_to!($src, $expected, $ty, $transform, false);
$crate::assert_wasm_evals_to!($src, $expected, $ty, $transform);
}
};
}

1
compiler/gen_wasm/tests/helpers/mod.rs
View file

@ -2,6 +2,7 @@ extern crate bumpalo;
#[macro_use]
pub mod eval;
pub mod wasm32_test_result;
/// Used in the with_larger_debug_stack() function, for tests that otherwise
/// run out of stack space in debug builds (but don't in --release builds)

278
compiler/gen_wasm/tests/helpers/wasm32_test_result.rs Normal file
View file

@ -0,0 +1,278 @@
use parity_wasm::builder;
use parity_wasm::builder::ModuleBuilder;
use parity_wasm::elements::{
Instruction, Instruction::*, Instructions, Internal, Local, ValueType,
};
use roc_gen_wasm::from_wasm32_memory::FromWasm32Memory;
use roc_gen_wasm::*;
use roc_std::{RocDec, RocList, RocOrder, RocStr};
const STACK_POINTER_LOCAL_ID: u32 = 0;
pub trait Wasm32TestResult {
fn insert_test_wrapper(
module_builder: &mut ModuleBuilder,
wrapper_name: &str,
main_function_index: u32,
) {
let instructions = Self::build_wrapper_body(main_function_index);
let signature = builder::signature().with_result(ValueType::I32).build_sig();
let stack_frame_pointer = Local::new(1, ValueType::I32);
let function_def = builder::function()
.with_signature(signature)
.body()
.with_locals(vec![stack_frame_pointer])
.with_instructions(Instructions::new(instructions))
.build() // body
.build(); // function
let location = module_builder.push_function(function_def);
let export = builder::export()
.field(wrapper_name)
.with_internal(Internal::Function(location.body))
.build();
module_builder.push_export(export);
}
fn build_wrapper_body(main_function_index: u32) -> Vec<Instruction>;
}
macro_rules! build_wrapper_body_primitive {
($store_instruction: expr, $align: expr) => {
fn build_wrapper_body(main_function_index: u32) -> Vec<Instruction> {
let size: i32 = 8;
let mut instructions = Vec::with_capacity(16);
push_stack_frame(&mut instructions, size, LocalId(STACK_POINTER_LOCAL_ID));
instructions.extend([
// load result address to prepare for the store instruction later
GetLocal(STACK_POINTER_LOCAL_ID),
//
// Call the main function with no arguments. Get primitive back.
Call(main_function_index),
//
// Store the primitive at the allocated address
$store_instruction($align, 0),
//
// Return the result pointer
GetLocal(STACK_POINTER_LOCAL_ID),
]);
pop_stack_frame(&mut instructions, size, LocalId(STACK_POINTER_LOCAL_ID));
instructions.push(End);
instructions
}
};
}
macro_rules! wasm_test_result_primitive {
($type_name: ident, $store_instruction: expr, $align: expr) => {
impl Wasm32TestResult for $type_name {
build_wrapper_body_primitive!($store_instruction, $align);
}
};
}
fn build_wrapper_body_stack_memory(main_function_index: u32, size: usize) -> Vec<Instruction> {
let mut instructions = Vec::with_capacity(16);
push_stack_frame(
&mut instructions,
size as i32,
LocalId(STACK_POINTER_LOCAL_ID),
);
instructions.extend([
//
// Call the main function with the allocated address to write the result.
// No value is returned to the VM stack. This is the same as in compiled C.
GetLocal(STACK_POINTER_LOCAL_ID),
Call(main_function_index),
//
// Return the result address
GetLocal(STACK_POINTER_LOCAL_ID),
]);
pop_stack_frame(
&mut instructions,
size as i32,
LocalId(STACK_POINTER_LOCAL_ID),
);
instructions.push(End);
instructions
}
macro_rules! wasm_test_result_stack_memory {
($type_name: ident) => {
impl Wasm32TestResult for $type_name {
fn build_wrapper_body(main_function_index: u32) -> Vec<Instruction> {
build_wrapper_body_stack_memory(main_function_index, $type_name::ACTUAL_WIDTH)
}
}
};
}
wasm_test_result_primitive!(bool, I32Store8, ALIGN_1);
wasm_test_result_primitive!(RocOrder, I32Store8, ALIGN_1);
wasm_test_result_primitive!(u8, I32Store8, ALIGN_1);
wasm_test_result_primitive!(i8, I32Store8, ALIGN_1);
wasm_test_result_primitive!(u16, I32Store16, ALIGN_2);
wasm_test_result_primitive!(i16, I32Store16, ALIGN_2);
wasm_test_result_primitive!(u32, I32Store, ALIGN_4);
wasm_test_result_primitive!(i32, I32Store, ALIGN_4);
wasm_test_result_primitive!(u64, I64Store, ALIGN_8);
wasm_test_result_primitive!(i64, I64Store, ALIGN_8);
wasm_test_result_primitive!(f32, F32Store, ALIGN_8);
wasm_test_result_primitive!(f64, F64Store, ALIGN_8);
wasm_test_result_stack_memory!(u128);
wasm_test_result_stack_memory!(i128);
wasm_test_result_stack_memory!(RocDec);
wasm_test_result_stack_memory!(RocStr);
impl<T: Wasm32TestResult> Wasm32TestResult for RocList<T> {
fn build_wrapper_body(main_function_index: u32) -> Vec<Instruction> {
build_wrapper_body_stack_memory(main_function_index, 12)
}
}
impl<T: Wasm32TestResult> Wasm32TestResult for &'_ T {
build_wrapper_body_primitive!(I32Store, ALIGN_4);
}
impl<T, const N: usize> Wasm32TestResult for [T; N]
where
T: Wasm32TestResult + FromWasm32Memory,
{
fn build_wrapper_body(main_function_index: u32) -> Vec<Instruction> {
build_wrapper_body_stack_memory(main_function_index, N * T::ACTUAL_WIDTH)
}
}
impl<T, U> Wasm32TestResult for (T, U)
where
T: Wasm32TestResult + FromWasm32Memory,
U: Wasm32TestResult + FromWasm32Memory,
{
fn build_wrapper_body(main_function_index: u32) -> Vec<Instruction> {
build_wrapper_body_stack_memory(main_function_index, T::ACTUAL_WIDTH + U::ACTUAL_WIDTH)
}
}
impl<T, U, V> Wasm32TestResult for (T, U, V)
where
T: Wasm32TestResult + FromWasm32Memory,
U: Wasm32TestResult + FromWasm32Memory,
V: Wasm32TestResult + FromWasm32Memory,
{
fn build_wrapper_body(main_function_index: u32) -> Vec<Instruction> {
build_wrapper_body_stack_memory(
main_function_index,
T::ACTUAL_WIDTH + U::ACTUAL_WIDTH + V::ACTUAL_WIDTH,
)
}
}
impl<T, U, V, W> Wasm32TestResult for (T, U, V, W)
where
T: Wasm32TestResult + FromWasm32Memory,
U: Wasm32TestResult + FromWasm32Memory,
V: Wasm32TestResult + FromWasm32Memory,
W: Wasm32TestResult + FromWasm32Memory,
{
fn build_wrapper_body(main_function_index: u32) -> Vec<Instruction> {
build_wrapper_body_stack_memory(
main_function_index,
T::ACTUAL_WIDTH + U::ACTUAL_WIDTH + V::ACTUAL_WIDTH + W::ACTUAL_WIDTH,
)
}
}
impl<T, U, V, W, X> Wasm32TestResult for (T, U, V, W, X)
where
T: Wasm32TestResult + FromWasm32Memory,
U: Wasm32TestResult + FromWasm32Memory,
V: Wasm32TestResult + FromWasm32Memory,
W: Wasm32TestResult + FromWasm32Memory,
X: Wasm32TestResult + FromWasm32Memory,
{
fn build_wrapper_body(main_function_index: u32) -> Vec<Instruction> {
build_wrapper_body_stack_memory(
main_function_index,
T::ACTUAL_WIDTH + U::ACTUAL_WIDTH + V::ACTUAL_WIDTH + W::ACTUAL_WIDTH + X::ACTUAL_WIDTH,
)
}
}
impl<T, U, V, W, X, Y> Wasm32TestResult for (T, U, V, W, X, Y)
where
T: Wasm32TestResult + FromWasm32Memory,
U: Wasm32TestResult + FromWasm32Memory,
V: Wasm32TestResult + FromWasm32Memory,
W: Wasm32TestResult + FromWasm32Memory,
X: Wasm32TestResult + FromWasm32Memory,
Y: Wasm32TestResult + FromWasm32Memory,
{
fn build_wrapper_body(main_function_index: u32) -> Vec<Instruction> {
build_wrapper_body_stack_memory(
main_function_index,
T::ACTUAL_WIDTH
+ U::ACTUAL_WIDTH
+ V::ACTUAL_WIDTH
+ W::ACTUAL_WIDTH
+ X::ACTUAL_WIDTH
+ Y::ACTUAL_WIDTH,
)
}
}
impl<T, U, V, W, X, Y, Z> Wasm32TestResult for (T, U, V, W, X, Y, Z)
where
T: Wasm32TestResult + FromWasm32Memory,
U: Wasm32TestResult + FromWasm32Memory,
V: Wasm32TestResult + FromWasm32Memory,
W: Wasm32TestResult + FromWasm32Memory,
X: Wasm32TestResult + FromWasm32Memory,
Y: Wasm32TestResult + FromWasm32Memory,
Z: Wasm32TestResult + FromWasm32Memory,
{
fn build_wrapper_body(main_function_index: u32) -> Vec<Instruction> {
build_wrapper_body_stack_memory(
main_function_index,
T::ACTUAL_WIDTH
+ U::ACTUAL_WIDTH
+ V::ACTUAL_WIDTH
+ W::ACTUAL_WIDTH
+ X::ACTUAL_WIDTH
+ Y::ACTUAL_WIDTH
+ Z::ACTUAL_WIDTH,
)
}
}
impl<T, U, V, W, X, Y, Z, A> Wasm32TestResult for (T, U, V, W, X, Y, Z, A)
where
T: Wasm32TestResult + FromWasm32Memory,
U: Wasm32TestResult + FromWasm32Memory,
V: Wasm32TestResult + FromWasm32Memory,
W: Wasm32TestResult + FromWasm32Memory,
X: Wasm32TestResult + FromWasm32Memory,
Y: Wasm32TestResult + FromWasm32Memory,
Z: Wasm32TestResult + FromWasm32Memory,
A: Wasm32TestResult + FromWasm32Memory,
{
fn build_wrapper_body(main_function_index: u32) -> Vec<Instruction> {
build_wrapper_body_stack_memory(
main_function_index,
T::ACTUAL_WIDTH
+ U::ACTUAL_WIDTH
+ V::ACTUAL_WIDTH
+ W::ACTUAL_WIDTH
+ X::ACTUAL_WIDTH
+ Y::ACTUAL_WIDTH
+ Z::ACTUAL_WIDTH
+ A::ACTUAL_WIDTH,
)
}
}

241
compiler/gen_wasm/tests/wasm_num.rs
View file

@ -11,7 +11,7 @@ extern crate libc;
mod helpers;
#[cfg(all(test, any(target_os = "linux", target_os = "macos"), any(target_arch = "x86_64"/*, target_arch = "aarch64"*/)))]
mod dev_num {
mod wasm_num {
#[test]
fn i64_values() {
assert_evals_to!("0", 0, i64);
@ -36,6 +36,101 @@ mod dev_num {
assert_evals_to!(&format!("{:0.1}", f64::MAX), f64::MAX, f64);
}
#[test]
fn i8_add_wrap() {
assert_evals_to!(
indoc!(
r#"
x : I8
x = 0x7f + 0x7f
x
"#
),
-2,
i8
);
}
#[test]
fn i16_add_wrap() {
assert_evals_to!(
indoc!(
r#"
x : I16
x = 0x7fff + 0x7fff
x
"#
),
-2,
i16
);
}
#[test]
fn i32_add_wrap() {
assert_evals_to!(
indoc!(
r#"
x : I32
x = 0x7fffffff + 0x7fffffff
x
"#
),
-2,
i32
);
}
#[test]
fn u8_add_wrap() {
assert_evals_to!(
indoc!(
r#"
x : U8
x = 0xff + 0xff
x
"#
),
0xfe,
u8
);
}
#[test]
fn u16_add_wrap() {
assert_evals_to!(
indoc!(
r#"
x : U16
x = 0xffff + 0xffff
x
"#
),
0xfffe,
u16
);
}
#[test]
fn u32_add_wrap() {
assert_evals_to!(
indoc!(
r#"
x : U32
x = 0xffffffff + 0xffffffff
x
"#
),
0xfffffffe,
u32
);
}
#[test]
fn gen_add_i64() {
assert_evals_to!(
@ -154,44 +249,44 @@ mod dev_num {
);
}
// #[test]
// fn gen_add_f64() {
// assert_evals_to!(
// indoc!(
// r#"
// 1.1 + 2.4 + 3
// "#
// ),
// 6.5,
// f64
// );
// }
#[test]
fn gen_add_f64() {
assert_evals_to!(
indoc!(
r#"
1.1 + 2.4 + 3
"#
),
6.5,
f64
);
}
// #[test]
// fn gen_sub_i64() {
// assert_evals_to!(
// indoc!(
// r#"
// 1 - 2 - 3
// "#
// ),
// -4,
// i64
// );
// }
#[test]
fn gen_sub_i64() {
assert_evals_to!(
indoc!(
r#"
1 - 2 - 3
"#
),
-4,
i64
);
}
// #[test]
// fn gen_mul_i64() {
// assert_evals_to!(
// indoc!(
// r#"
// 2 * 4 * 6
// "#
// ),
// 48,
// i64
// );
// }
#[test]
fn gen_mul_i64() {
assert_evals_to!(
indoc!(
r#"
2 * 4 * 6
"#
),
48,
i64
);
}
#[test]
fn i64_force_stack() {
@ -476,18 +571,18 @@ mod dev_num {
// );
// }
// #[test]
// fn gen_sub_f64() {
// assert_evals_to!(
// indoc!(
// r#"
// 1.5 - 2.4 - 3
// "#
// ),
// -3.9,
// f64
// );
// }
#[test]
fn gen_sub_f64() {
assert_evals_to!(
indoc!(
r#"
1.5 - 2.4 - 3
"#
),
-3.9,
f64
);
}
// #[test]
// fn gen_div_i64() {
@ -685,31 +780,31 @@ mod dev_num {
// assert_evals_to!("0.0 >= 0.0", true, bool);
// }
// #[test]
// fn gen_order_of_arithmetic_ops() {
// assert_evals_to!(
// indoc!(
// r#"
// 1 + 3 * 7 - 2
// "#
// ),
// 20,
// i64
// );
// }
#[test]
fn gen_order_of_arithmetic_ops() {
assert_evals_to!(
indoc!(
r#"
1 + 3 * 7 - 2
"#
),
20,
i64
);
}
// #[test]
// fn gen_order_of_arithmetic_ops_complex_float() {
// assert_evals_to!(
// indoc!(
// r#"
// 3 - 48 * 2.0
// "#
// ),
// -93.0,
// f64
// );
// }
#[test]
fn gen_order_of_arithmetic_ops_complex_float() {
assert_evals_to!(
indoc!(
r#"
3 - 48 * 2.0
"#
),
-93.0,
f64
);
}
// #[test]
// fn if_guard_bind_variable_false() {

512
compiler/gen_wasm/tests/wasm_records.rs
View file

@ -5,7 +5,7 @@ extern crate indoc;
mod helpers;
#[cfg(all(test, target_os = "linux", any(target_arch = "x86_64"/*, target_arch = "aarch64"*/)))]
mod dev_records {
mod wasm_records {
// #[test]
// fn basic_record() {
// assert_evals_to!(
@ -307,142 +307,116 @@ mod dev_records {
// ()
// );
// }
//
// #[test]
// fn i64_record1_literal() {
// assert_evals_to!(
// indoc!(
// r#"
// { x: 3 }
// "#
// ),
// 3,
// i64
// );
// }
// #[test]
// fn i64_record2_literal() {
// assert_evals_to!(
// indoc!(
// r#"
// { x: 3, y: 5 }
// "#
// ),
// (3, 5),
// (i64, i64)
// );
// }
#[test]
fn i64_record1_literal() {
assert_evals_to!(
indoc!(
r#"
{ x: 3 }
"#
),
3,
i64
);
}
// // #[test]
// // fn i64_record3_literal() {
// // assert_evals_to!(
// // indoc!(
// // r#"
// // { x: 3, y: 5, z: 17 }
// // "#
// // ),
// // (3, 5, 17),
// // (i64, i64, i64)
// // );
// // }
#[test]
fn i64_record2_literal() {
assert_evals_to!(
indoc!(
r#"
{ x: 3, y: 5 }
"#
),
(3, 5),
(i64, i64)
);
}
// #[test]
// fn f64_record2_literal() {
// assert_evals_to!(
// indoc!(
// r#"
// { x: 3.1, y: 5.1 }
// "#
// ),
// (3.1, 5.1),
// (f64, f64)
// );
// }
#[test]
fn i64_record3_literal() {
assert_evals_to!(
indoc!(
r#"
{ x: 3, y: 5, z: 17 }
"#
),
(3, 5, 17),
(i64, i64, i64)
);
}
// // #[test]
// // fn f64_record3_literal() {
// // assert_evals_to!(
// // indoc!(
// // r#"
// // { x: 3.1, y: 5.1, z: 17.1 }
// // "#
// // ),
// // (3.1, 5.1, 17.1),
// // (f64, f64, f64)
// // );
// // }
#[test]
fn f64_record2_literal() {
assert_evals_to!(
indoc!(
r#"
{ x: 3.1, y: 5.1 }
"#
),
(3.1, 5.1),
(f64, f64)
);
}
// // #[test]
// // fn bool_record4_literal() {
// // assert_evals_to!(
// // indoc!(
// // r#"
// // record : { a : Bool, b : Bool, c : Bool, d : Bool }
// // record = { a: True, b: True, c : True, d : Bool }
#[test]
fn f64_record3_literal() {
assert_evals_to!(
indoc!(
r#"
{ x: 3.1, y: 5.1, z: 17.1 }
"#
),
(3.1, 5.1, 17.1),
(f64, f64, f64)
);
}
// // record
// // "#
// // ),
// // (true, false, false, true),
// // (bool, bool, bool, bool)
// // );
// // }
#[test]
fn bool_record4_literal() {
assert_evals_to!(
indoc!(
r#"
record : { a : Bool, b : Bool, c : Bool, d : Bool }
record = { a: True, b: False, c : False, d : True }
// #[test]
// fn i64_record1_literal() {
// assert_evals_to!(
// indoc!(
// r#"
// { a: 3 }
// "#
// ),
// 3,
// i64
// );
// }
record
"#
),
[true, false, false, true],
[bool; 4]
);
}
// // #[test]
// // fn i64_record9_literal() {
// // assert_evals_to!(
// // indoc!(
// // r#"
// // { a: 3, b: 5, c: 17, d: 1, e: 9, f: 12, g: 13, h: 14, i: 15 }
// // "#
// // ),
// // (3, 5, 17, 1, 9, 12, 13, 14, 15),
// // (i64, i64, i64, i64, i64, i64, i64, i64, i64)
// // );
// // }
#[test]
fn i64_record9_literal() {
assert_evals_to!(
indoc!(
r#"
{ a: 3, b: 5, c: 17, d: 1, e: 9, f: 12, g: 13, h: 14, i: 15 }
"#
),
[3, 5, 17, 1, 9, 12, 13, 14, 15],
[i64; 9]
);
}
// // #[test]
// // fn f64_record3_literal() {
// // assert_evals_to!(
// // indoc!(
// // r#"
// // { x: 3.1, y: 5.1, z: 17.1 }
// // "#
// // ),
// // (3.1, 5.1, 17.1),
// // (f64, f64, f64)
// // );
// // }
#[test]
fn bool_literal() {
assert_evals_to!(
indoc!(
r#"
x : Bool
x = True
// #[test]
// fn bool_literal() {
// assert_evals_to!(
// indoc!(
// r#"
// x : Bool
// x = True
// x
// "#
// ),
// true,
// bool
// );
// }
x
"#
),
true,
bool
);
}
// #[test]
// fn optional_field_when_use_default() {
@ -667,135 +641,135 @@ mod dev_records {
// );
// }
// #[test]
// fn return_record_2() {
// assert_evals_to!(
// indoc!(
// r#"
// { x: 3, y: 5 }
// "#
// ),
// [3, 5],
// [i64; 2]
// );
// }
#[test]
fn return_record_2() {
assert_evals_to!(
indoc!(
r#"
{ x: 3, y: 5 }
"#
),
[3, 5],
[i64; 2]
);
}
// #[test]
// fn return_record_3() {
// assert_evals_to!(
// indoc!(
// r#"
// { x: 3, y: 5, z: 4 }
// "#
// ),
// (3, 5, 4),
// (i64, i64, i64)
// );
// }
#[test]
fn return_record_3() {
assert_evals_to!(
indoc!(
r#"
{ x: 3, y: 5, z: 4 }
"#
),
(3, 5, 4),
(i64, i64, i64)
);
}
// #[test]
// fn return_record_4() {
// assert_evals_to!(
// indoc!(
// r#"
// { a: 3, b: 5, c: 4, d: 2 }
// "#
// ),
// [3, 5, 4, 2],
// [i64; 4]
// );
// }
#[test]
fn return_record_4() {
assert_evals_to!(
indoc!(
r#"
{ a: 3, b: 5, c: 4, d: 2 }
"#
),
[3, 5, 4, 2],
[i64; 4]
);
}
// #[test]
// fn return_record_5() {
// assert_evals_to!(
// indoc!(
// r#"
// { a: 3, b: 5, c: 4, d: 2, e: 1 }
// "#
// ),
// [3, 5, 4, 2, 1],
// [i64; 5]
// );
// }
#[test]
fn return_record_5() {
assert_evals_to!(
indoc!(
r#"
{ a: 3, b: 5, c: 4, d: 2, e: 1 }
"#
),
[3, 5, 4, 2, 1],
[i64; 5]
);
}
// #[test]
// fn return_record_6() {
// assert_evals_to!(
// indoc!(
// r#"
// { a: 3, b: 5, c: 4, d: 2, e: 1, f: 7 }
// "#
// ),
// [3, 5, 4, 2, 1, 7],
// [i64; 6]
// );
// }
#[test]
fn return_record_6() {
assert_evals_to!(
indoc!(
r#"
{ a: 3, b: 5, c: 4, d: 2, e: 1, f: 7 }
"#
),
[3, 5, 4, 2, 1, 7],
[i64; 6]
);
}
// #[test]
// fn return_record_7() {
// assert_evals_to!(
// indoc!(
// r#"
// { a: 3, b: 5, c: 4, d: 2, e: 1, f: 7, g: 8 }
// "#
// ),
// [3, 5, 4, 2, 1, 7, 8],
// [i64; 7]
// );
// }
#[test]
fn return_record_7() {
assert_evals_to!(
indoc!(
r#"
{ a: 3, b: 5, c: 4, d: 2, e: 1, f: 7, g: 8 }
"#
),
[3, 5, 4, 2, 1, 7, 8],
[i64; 7]
);
}
// #[test]
// fn return_record_float_int() {
// assert_evals_to!(
// indoc!(
// r#"
// { a: 3.14, b: 0x1 }
// "#
// ),
// (3.14, 0x1),
// (f64, i64)
// );
// }
#[test]
fn return_record_float_int() {
assert_evals_to!(
indoc!(
r#"
{ a: 3.14, b: 0x1 }
"#
),
(3.14, 0x1),
(f64, i64)
);
}
// #[test]
// fn return_record_int_float() {
// assert_evals_to!(
// indoc!(
// r#"
// { a: 0x1, b: 3.14 }
// "#
// ),
// (0x1, 3.14),
// (i64, f64)
// );
// }
#[test]
fn return_record_int_float() {
assert_evals_to!(
indoc!(
r#"
{ a: 0x1, b: 3.14 }
"#
),
(0x1, 3.14),
(i64, f64)
);
}
// #[test]
// fn return_record_float_float() {
// assert_evals_to!(
// indoc!(
// r#"
// { a: 6.28, b: 3.14 }
// "#
// ),
// (6.28, 3.14),
// (f64, f64)
// );
// }
#[test]
fn return_record_float_float() {
assert_evals_to!(
indoc!(
r#"
{ a: 6.28, b: 3.14 }
"#
),
(6.28, 3.14),
(f64, f64)
);
}
// #[test]
// fn return_record_float_float_float() {
// assert_evals_to!(
// indoc!(
// r#"
// { a: 6.28, b: 3.14, c: 0.1 }
// "#
// ),
// (6.28, 3.14, 0.1),
// (f64, f64, f64)
// );
// }
#[test]
fn return_record_float_float_float() {
assert_evals_to!(
indoc!(
r#"
{ a: 6.28, b: 3.14, c: 0.1 }
"#
),
(6.28, 3.14, 0.1),
(f64, f64, f64)
);
}
// #[test]
// fn return_nested_record() {
@ -851,20 +825,20 @@ mod dev_records {
// );
// }
#[test]
fn update_single_element_record() {
assert_evals_to!(
indoc!(
r#"
rec = { foo: 42}
// #[test]
// fn update_single_element_record() {
// assert_evals_to!(
// indoc!(
// r#"
// rec = { foo: 42}
{ rec & foo: rec.foo + 1 }
"#
),
43,
i64
);
}
// { rec & foo: rec.foo + 1 }
// "#
// ),
// 43,
// i64
// );
// }
// #[test]
// fn booleans_in_record() {
@ -899,6 +873,24 @@ mod dev_records {
// );
// }
#[test]
fn stack_memory_return_from_branch() {
// stack memory pointer should end up in the right place after returning from a branch
assert_evals_to!(
indoc!(
r#"
stackMemoryJunk = { x: 999, y: 111 }
if True then
{ x: 123, y: 321 }
else
stackMemoryJunk
"#
),
(123, 321),
(i64, i64)
);
}
// #[test]
// fn blue_and_present() {
// assert_evals_to!(

13
compiler/load/src/file.rs
View file

@ -37,13 +37,13 @@ use roc_types::subs::{Subs, VarStore, Variable};
use roc_types::types::{Alias, Type};
use std::collections::hash_map::Entry::{Occupied, Vacant};
use std::collections::{HashMap, HashSet};
use std::fs;
use std::io;
use std::iter;
use std::path::{Path, PathBuf};
use std::str::from_utf8_unchecked;
use std::sync::Arc;
use std::time::{Duration, SystemTime};
use std::{env, fs};
/// Default name for the binary generated for an app, if an invalid one was specified.
const DEFAULT_APP_OUTPUT_PATH: &str = "app";
@ -1351,7 +1351,12 @@ where
// doing .max(1) on the entire expression guards against
// num_cpus returning 0, while also avoiding wrapping
// unsigned subtraction overflow.
let num_workers = num_cpus::get().max(2) - 1;
let default_num_workers = num_cpus::get().max(2) - 1;
let num_workers = match env::var("ROC_NUM_WORKERS") {
Ok(env_str) => env_str.parse::<usize>().unwrap_or(default_num_workers),
Err(_) => default_num_workers,
};
let worker_arenas = arena.alloc(bumpalo::collections::Vec::with_capacity_in(
num_workers,
@ -2113,8 +2118,6 @@ fn update<'a>(
&mut state.procedures,
);
Proc::insert_refcount_operations(arena, &mut state.procedures);
// display the mono IR of the module, for debug purposes
if roc_mono::ir::PRETTY_PRINT_IR_SYMBOLS {
let procs_string = state
@ -2128,6 +2131,8 @@ fn update<'a>(
println!("{}", result);
}
Proc::insert_refcount_operations(arena, &mut state.procedures);
// This is not safe with the new non-recursive RC updates that we do for tag unions
//
// Proc::optimize_refcount_operations(

27
compiler/module/src/low_level.rs
View file

@ -15,6 +15,7 @@ pub enum LowLevel {
StrFromUtf8,
StrFromUtf8Range,
StrToUtf8,
StrRepeat,
StrFromFloat,
ListLen,
ListGetUnsafe,
@ -114,19 +115,19 @@ impl LowLevel {
match self {
StrConcat | StrJoinWith | StrIsEmpty | StrStartsWith | StrStartsWithCodePt
| StrEndsWith | StrSplit | StrCountGraphemes | StrFromInt | StrFromUtf8
| StrFromUtf8Range | StrToUtf8 | StrFromFloat | ListLen | ListGetUnsafe | ListSet
| ListDrop | ListSingle | ListRepeat | ListReverse | ListConcat | ListContains
| ListAppend | ListPrepend | ListJoin | ListRange | ListSwap | DictSize | DictEmpty
| DictInsert | DictRemove | DictContains | DictGetUnsafe | DictKeys | DictValues
| DictUnion | DictIntersection | DictDifference | SetFromList | NumAdd | NumAddWrap
| NumAddChecked | NumSub | NumSubWrap | NumSubChecked | NumMul | NumMulWrap
| NumMulChecked | NumGt | NumGte | NumLt | NumLte | NumCompare | NumDivUnchecked
| NumRemUnchecked | NumIsMultipleOf | NumAbs | NumNeg | NumSin | NumCos
| NumSqrtUnchecked | NumLogUnchecked | NumRound | NumToFloat | NumPow | NumCeiling
| NumPowInt | NumFloor | NumIsFinite | NumAtan | NumAcos | NumAsin | NumBitwiseAnd
| NumBitwiseXor | NumBitwiseOr | NumShiftLeftBy | NumShiftRightBy | NumBytesToU16
| NumBytesToU32 | NumShiftRightZfBy | NumIntCast | Eq | NotEq | And | Or | Not
| Hash | ExpectTrue => false,
| StrFromUtf8Range | StrToUtf8 | StrRepeat | StrFromFloat | ListLen | ListGetUnsafe
| ListSet | ListDrop | ListSingle | ListRepeat | ListReverse | ListConcat
| ListContains | ListAppend | ListPrepend | ListJoin | ListRange | ListSwap
| DictSize | DictEmpty | DictInsert | DictRemove | DictContains | DictGetUnsafe
| DictKeys | DictValues | DictUnion | DictIntersection | DictDifference
| SetFromList | NumAdd | NumAddWrap | NumAddChecked | NumSub | NumSubWrap
| NumSubChecked | NumMul | NumMulWrap | NumMulChecked | NumGt | NumGte | NumLt
| NumLte | NumCompare | NumDivUnchecked | NumRemUnchecked | NumIsMultipleOf
| NumAbs | NumNeg | NumSin | NumCos | NumSqrtUnchecked | NumLogUnchecked | NumRound
| NumToFloat | NumPow | NumCeiling | NumPowInt | NumFloor | NumIsFinite | NumAtan
| NumAcos | NumAsin | NumBitwiseAnd | NumBitwiseXor | NumBitwiseOr | NumShiftLeftBy
| NumShiftRightBy | NumBytesToU16 | NumBytesToU32 | NumShiftRightZfBy | NumIntCast
| Eq | NotEq | And | Or | Not | Hash | ExpectTrue => false,
ListMap | ListMap2 | ListMap3 | ListMapWithIndex | ListKeepIf | ListWalk
| ListWalkUntil | ListWalkBackwards | ListKeepOks | ListKeepErrs | ListSortWith

1
compiler/module/src/symbol.rs
View file

@ -927,6 +927,7 @@ define_builtins! {
16 STR_STARTS_WITH_CODE_PT: "startsWithCodePt"
17 STR_ALIAS_ANALYSIS_STATIC: "#aliasAnalysisStatic" // string with the static lifetime
18 STR_FROM_UTF8_RANGE: "fromUtf8Range"
19 STR_REPEAT: "repeat"
}
4 LIST: "List" => {
0 LIST_LIST: "List" imported // the List.List type alias

52
compiler/mono/src/alias_analysis.rs
View file

@ -16,6 +16,7 @@ use crate::layout::{Builtin, Layout, ListLayout, UnionLayout};
pub const MOD_APP: ModName = ModName(b"UserApp");
pub const STATIC_STR_NAME: ConstName = ConstName(&Symbol::STR_ALIAS_ANALYSIS_STATIC.to_ne_bytes());
pub const STATIC_LIST_NAME: ConstName = ConstName(b"THIS IS A STATIC LIST");
const ENTRY_POINT_NAME: &[u8] = b"mainForHost";
@ -128,6 +129,22 @@ where
};
m.add_const(STATIC_STR_NAME, static_str_def)?;
// a const that models all static lists
let static_list_def = {
let mut cbuilder = ConstDefBuilder::new();
let block = cbuilder.add_block();
let cell = cbuilder.add_new_heap_cell(block)?;
let unit_type = cbuilder.add_tuple_type(&[])?;
let bag = cbuilder.add_empty_bag(block, unit_type)?;
let value_id = cbuilder.add_make_tuple(block, &[cell, bag])?;
let root = BlockExpr(block, value_id);
let list_type_id = static_list_type(&mut cbuilder)?;
cbuilder.build(list_type_id, root)?
};
m.add_const(STATIC_LIST_NAME, static_list_def)?;
// the entry point wrapper
let roc_main_bytes = func_name_bytes_help(
entry_point.symbol,
@ -589,9 +606,9 @@ fn call_spec(
let index = builder.add_make_tuple(block, &[])?;
let argument = if closure_env_layout.is_none() {
builder.add_make_tuple(block, &[first, index])?
builder.add_make_tuple(block, &[index, first])?
} else {
builder.add_make_tuple(block, &[first, index, closure_env])?
builder.add_make_tuple(block, &[index, first, closure_env])?
};
builder.add_call(block, spec_var, module, name, argument)?;
}
@ -1117,9 +1134,11 @@ fn expr_spec<'a>(
let list = new_list(builder, block, type_id)?;
let mut bag = builder.add_get_tuple_field(block, list, LIST_BAG_INDEX)?;
let mut all_constants = true;
for element in elems.iter() {
let value_id = if let ListLiteralElement::Symbol(symbol) = element {
all_constants = false;
env.symbols[symbol]
} else {
builder.add_make_tuple(block, &[]).unwrap()
@ -1128,9 +1147,13 @@ fn expr_spec<'a>(
bag = builder.add_bag_insert(block, bag, value_id)?;
}
let cell = builder.add_new_heap_cell(block)?;
if all_constants {
new_static_list(builder, block)
} else {
let cell = builder.add_new_heap_cell(block)?;
builder.add_make_tuple(block, &[cell, bag])
builder.add_make_tuple(block, &[cell, bag])
}
}
EmptyArray => {
@ -1191,6 +1214,11 @@ fn layout_spec_help(
match layout {
Builtin(builtin) => builtin_spec(builder, builtin, when_recursive),
Struct(fields) => build_recursive_tuple_type(builder, fields, when_recursive),
LambdaSet(lambda_set) => layout_spec_help(
builder,
&lambda_set.runtime_representation(),
when_recursive,
),
Union(union_layout) => {
let variant_types = build_variant_types(builder, union_layout)?;
@ -1236,7 +1264,7 @@ fn builtin_spec(
match builtin {
Int128 | Int64 | Int32 | Int16 | Int8 | Int1 | Usize => builder.add_tuple_type(&[]),
Decimal | Float128 | Float64 | Float32 | Float16 => builder.add_tuple_type(&[]),
Decimal | Float128 | Float64 | Float32 => builder.add_tuple_type(&[]),
Str | EmptyStr => str_type(builder),
Dict(key_layout, value_layout) => {
let value_type = layout_spec_help(builder, value_layout, when_recursive)?;
@ -1291,6 +1319,14 @@ fn str_type<TC: TypeContext>(builder: &mut TC) -> Result<TypeId> {
builder.add_tuple_type(&[cell_id])
}
fn static_list_type<TC: TypeContext>(builder: &mut TC) -> Result<TypeId> {
let unit_type = builder.add_tuple_type(&[])?;
let cell = builder.add_heap_cell_type();
let bag = builder.add_bag_type(unit_type)?;
builder.add_tuple_type(&[cell, bag])
}
// const OK_TAG_ID: u8 = 1u8;
// const ERR_TAG_ID: u8 = 0u8;
@ -1324,6 +1360,12 @@ fn new_static_string(builder: &mut FuncDefBuilder, block: BlockId) -> Result<Val
builder.add_const_ref(block, module, STATIC_STR_NAME)
}
fn new_static_list(builder: &mut FuncDefBuilder, block: BlockId) -> Result<ValueId> {
let module = MOD_APP;
builder.add_const_ref(block, module, STATIC_LIST_NAME)
}
fn new_num(builder: &mut FuncDefBuilder, block: BlockId) -> Result<ValueId> {
// we model all our numbers as unit values
builder.add_make_tuple(block, &[])

1
compiler/mono/src/borrow.rs
View file

@ -1013,6 +1013,7 @@ pub fn lowlevel_borrow_signature(arena: &Bump, op: LowLevel) -> &[bool] {
StrFromUtf8 => arena.alloc_slice_copy(&[owned]),
StrFromUtf8Range => arena.alloc_slice_copy(&[borrowed, irrelevant]),
StrToUtf8 => arena.alloc_slice_copy(&[owned]),
StrRepeat => arena.alloc_slice_copy(&[borrowed, irrelevant]),
StrFromInt | StrFromFloat => arena.alloc_slice_copy(&[irrelevant]),
Hash => arena.alloc_slice_copy(&[borrowed, irrelevant]),
DictSize => arena.alloc_slice_copy(&[borrowed]),

50
compiler/mono/src/decision_tree.rs
View file

@ -1,6 +1,7 @@
use crate::exhaustive::{Ctor, RenderAs, TagId, Union};
use crate::ir::{
BranchInfo, DestructType, Env, Expr, JoinPointId, Literal, Param, Pattern, Procs, Stmt,
BranchInfo, DestructType, Env, Expr, FloatPrecision, IntPrecision, JoinPointId, Literal, Param,
Pattern, Procs, Stmt,
};
use crate::layout::{Builtin, Layout, LayoutCache, UnionLayout};
use roc_collections::all::{MutMap, MutSet};
@ -85,8 +86,8 @@ enum Test<'a> {
union: crate::exhaustive::Union,
arguments: Vec<(Pattern<'a>, Layout<'a>)>,
},
IsInt(i128),
IsFloat(u64),
IsInt(i128, IntPrecision),
IsFloat(u64, FloatPrecision),
IsDecimal(RocDec),
IsStr(Box<str>),
IsBit(bool),
@ -95,6 +96,7 @@ enum Test<'a> {
num_alts: usize,
},
}
use std::hash::{Hash, Hasher};
impl<'a> Hash for Test<'a> {
fn hash<H: Hasher>(&self, state: &mut H) {
@ -106,13 +108,15 @@ impl<'a> Hash for Test<'a> {
tag_id.hash(state);
// The point of this custom implementation is to not hash the tag arguments
}
IsInt(v) => {
IsInt(v, width) => {
state.write_u8(1);
v.hash(state);
width.hash(state);
}
IsFloat(v) => {
IsFloat(v, width) => {
state.write_u8(2);
v.hash(state);
width.hash(state);
}
IsStr(v) => {
state.write_u8(3);
@ -306,8 +310,8 @@ fn tests_are_complete_help(last_test: &Test, number_of_tests: usize) -> bool {
Test::IsCtor { union, .. } => number_of_tests == union.alternatives.len(),
Test::IsByte { num_alts, .. } => number_of_tests == *num_alts,
Test::IsBit(_) => number_of_tests == 2,
Test::IsInt(_) => false,
Test::IsFloat(_) => false,
Test::IsInt(_, _) => false,
Test::IsFloat(_, _) => false,
Test::IsDecimal(_) => false,
Test::IsStr(_) => false,
}
@ -561,8 +565,8 @@ fn test_at_path<'a>(
tag_id: *tag_id,
num_alts: union.alternatives.len(),
},
IntLiteral(v) => IsInt(*v),
FloatLiteral(v) => IsFloat(*v),
IntLiteral(v, precision) => IsInt(*v, *precision),
FloatLiteral(v, precision) => IsFloat(*v, *precision),
DecimalLiteral(v) => IsDecimal(*v),
StrLiteral(v) => IsStr(v.clone()),
};
@ -807,8 +811,9 @@ fn to_relevant_branch_help<'a>(
_ => None,
},
IntLiteral(int) => match test {
IsInt(is_int) if int == *is_int => {
IntLiteral(int, p1) => match test {
IsInt(is_int, p2) if int == *is_int => {
debug_assert_eq!(p1, *p2);
start.extend(end);
Some(Branch {
goal: branch.goal,
@ -819,8 +824,9 @@ fn to_relevant_branch_help<'a>(
_ => None,
},
FloatLiteral(float) => match test {
IsFloat(test_float) if float == *test_float => {
FloatLiteral(float, p1) => match test {
IsFloat(test_float, p2) if float == *test_float => {
debug_assert_eq!(p1, *p2);
start.extend(end);
Some(Branch {
goal: branch.goal,
@ -928,8 +934,8 @@ fn needs_tests(pattern: &Pattern) -> bool {
| AppliedTag { .. }
| BitLiteral { .. }
| EnumLiteral { .. }
| IntLiteral(_)
| FloatLiteral(_)
| IntLiteral(_, _)
| FloatLiteral(_, _)
| DecimalLiteral(_)
| StrLiteral(_) => true,
}
@ -1280,22 +1286,22 @@ fn test_to_equality<'a>(
_ => unreachable!("{:?}", (cond_layout, union)),
}
}
Test::IsInt(test_int) => {
Test::IsInt(test_int, precision) => {
// TODO don't downcast i128 here
debug_assert!(test_int <= i64::MAX as i128);
let lhs = Expr::Literal(Literal::Int(test_int as i128));
let lhs_symbol = env.unique_symbol();
stores.push((lhs_symbol, Layout::Builtin(Builtin::Int64), lhs));
stores.push((lhs_symbol, precision.as_layout(), lhs));
(stores, lhs_symbol, rhs_symbol, None)
}
Test::IsFloat(test_int) => {
Test::IsFloat(test_int, precision) => {
// TODO maybe we can actually use i64 comparison here?
let test_float = f64::from_bits(test_int as u64);
let lhs = Expr::Literal(Literal::Float(test_float));
let lhs_symbol = env.unique_symbol();
stores.push((lhs_symbol, Layout::Builtin(Builtin::Float64), lhs));
stores.push((lhs_symbol, precision.as_layout(), lhs));
(stores, lhs_symbol, rhs_symbol, None)
}
@ -1303,7 +1309,7 @@ fn test_to_equality<'a>(
Test::IsDecimal(test_dec) => {
let lhs = Expr::Literal(Literal::Int(test_dec.0));
let lhs_symbol = env.unique_symbol();
stores.push((lhs_symbol, Layout::Builtin(Builtin::Int128), lhs));
stores.push((lhs_symbol, *cond_layout, lhs));
(stores, lhs_symbol, rhs_symbol, None)
}
@ -1737,8 +1743,8 @@ fn decide_to_branching<'a>(
);
let tag = match test {
Test::IsInt(v) => v as u64,
Test::IsFloat(v) => v as u64,
Test::IsInt(v, _) => v as u64,
Test::IsFloat(v, _) => v as u64,
Test::IsBit(v) => v as u64,
Test::IsByte { tag_id, .. } => tag_id as u64,
Test::IsCtor { tag_id, .. } => tag_id as u64,

4
compiler/mono/src/exhaustive.rs
View file

@ -65,8 +65,8 @@ fn simplify(pattern: &crate::ir::Pattern) -> Pattern {
use crate::ir::Pattern::*;
match pattern {
IntLiteral(v) => Literal(Literal::Int(*v)),
FloatLiteral(v) => Literal(Literal::Float(*v)),
IntLiteral(v, _) => Literal(Literal::Int(*v)),
FloatLiteral(v, _) => Literal(Literal::Float(*v)),
DecimalLiteral(v) => Literal(Literal::Decimal(*v)),
StrLiteral(v) => Literal(Literal::Str(v.clone())),

265
compiler/mono/src/ir.rs
View file

@ -54,6 +54,7 @@ macro_rules! return_on_layout_error_help {
#[derive(Debug, Clone, Copy)]
pub enum OptLevel {
Development,
Normal,
Optimize,
}
@ -694,7 +695,20 @@ impl<'a> Procs<'a> {
// the `layout` is a function pointer, while `_ignore_layout` can be a
// closure. We only specialize functions, storing this value with a closure
// layout will give trouble.
self.specialized.insert((symbol, layout), Done(proc));
let arguments =
Vec::from_iter_in(proc.args.iter().map(|(l, _)| *l), env.arena)
.into_bump_slice();
let proper_layout = ProcLayout {
arguments,
result: proc.ret_layout,
};
// NOTE: some function are specialized to have a closure, but don't actually
// need any closure argument. Here is where we correct this sort of thing,
// by trusting the layout of the Proc, not of what we specialize for
self.specialized.remove(&(symbol, layout));
self.specialized.insert((symbol, proper_layout), Done(proc));
}
Err(error) => {
panic!("TODO generate a RuntimeError message for {:?}", error);
@ -1843,7 +1857,7 @@ fn generate_runtime_error_function<'a>(
args.push((*arg, env.unique_symbol()));
}
args.push((lambda_set.runtime_representation(), Symbol::ARG_CLOSURE));
args.push((Layout::LambdaSet(lambda_set), Symbol::ARG_CLOSURE));
(args.into_bump_slice(), *ret_layout)
}
@ -1919,7 +1933,29 @@ fn specialize_external<'a>(
match layout {
RawFunctionLayout::Function(argument_layouts, lambda_set, return_layout) => {
let assigned = env.unique_symbol();
let unit = env.unique_symbol();
let mut argument_symbols =
Vec::with_capacity_in(argument_layouts.len(), env.arena);
let mut proc_arguments =
Vec::with_capacity_in(argument_layouts.len() + 1, env.arena);
let mut top_level_arguments =
Vec::with_capacity_in(argument_layouts.len() + 1, env.arena);
for layout in argument_layouts {
let symbol = env.unique_symbol();
proc_arguments.push((*layout, symbol));
argument_symbols.push(symbol);
top_level_arguments.push(*layout);
}
// the proc needs to take an extra closure argument
let lambda_set_layout = Layout::LambdaSet(lambda_set);
proc_arguments.push((lambda_set_layout, Symbol::ARG_CLOSURE));
// this should also be reflected in the TopLevel signature
top_level_arguments.push(lambda_set_layout);
let hole = env.arena.alloc(Stmt::Ret(assigned));
@ -1927,20 +1963,16 @@ fn specialize_external<'a>(
env,
lambda_set,
Symbol::ARG_CLOSURE,
env.arena.alloc([unit]),
argument_symbols.into_bump_slice(),
argument_layouts,
*return_layout,
assigned,
hole,
);
let body = let_empty_struct(unit, env.arena.alloc(body));
let proc = Proc {
name,
args: env
.arena
.alloc([(lambda_set.runtime_representation(), Symbol::ARG_CLOSURE)]),
args: proc_arguments.into_bump_slice(),
body,
closure_data_layout: None,
ret_layout: *return_layout,
@ -1951,7 +1983,7 @@ fn specialize_external<'a>(
let top_level = ProcLayout::new(
env.arena,
env.arena.alloc([lambda_set.runtime_representation()]),
top_level_arguments.into_bump_slice(),
*return_layout,
);
@ -1999,7 +2031,7 @@ fn specialize_external<'a>(
env.subs.rollback_to(snapshot);
let closure_data_layout = match opt_closure_layout {
Some(closure_layout) => closure_layout.runtime_representation(),
Some(lambda_set) => Layout::LambdaSet(lambda_set),
None => Layout::Struct(&[]),
};
@ -2149,7 +2181,7 @@ fn specialize_external<'a>(
env.subs.rollback_to(snapshot);
let closure_data_layout = match opt_closure_layout {
Some(closure_layout) => Some(closure_layout.runtime_representation()),
Some(lambda_set) => Some(Layout::LambdaSet(lambda_set)),
None => None,
};
@ -2260,7 +2292,7 @@ fn build_specialized_proc<'a>(
Some(lambda_set) if pattern_symbols.last() == Some(&Symbol::ARG_CLOSURE) => {
// here we define the lifted (now top-level) f function. Its final argument is `Symbol::ARG_CLOSURE`,
// it stores the closure structure (just an integer in this case)
proc_args.push((lambda_set.runtime_representation(), Symbol::ARG_CLOSURE));
proc_args.push((Layout::LambdaSet(lambda_set), Symbol::ARG_CLOSURE));
debug_assert_eq!(
pattern_layouts_len + 1,
@ -2297,7 +2329,7 @@ fn build_specialized_proc<'a>(
}
Ordering::Greater => {
if pattern_symbols.is_empty() {
let ret_layout = lambda_set.runtime_representation();
let ret_layout = Layout::LambdaSet(lambda_set);
Ok(FunctionPointerBody {
closure: None,
ret_layout,
@ -2482,7 +2514,7 @@ where
let raw = if procs.module_thunks.contains(&proc_name) {
match raw {
RawFunctionLayout::Function(_, lambda_set, _) => {
RawFunctionLayout::ZeroArgumentThunk(lambda_set.runtime_representation())
RawFunctionLayout::ZeroArgumentThunk(Layout::LambdaSet(lambda_set))
}
_ => raw,
}
@ -2656,6 +2688,7 @@ macro_rules! match_on_closure_argument {
let ret_layout = top_level.result;
match closure_data_layout {
RawFunctionLayout::Function(_, lambda_set, _) => {
lowlevel_match_on_lambda_set(
@ -2760,13 +2793,13 @@ pub fn with_hole<'a>(
IntOrFloat::SignedIntType(precision) => Stmt::Let(
assigned,
Expr::Literal(Literal::Int(int)),
Layout::Builtin(int_precision_to_builtin(precision)),
precision.as_layout(),
hole,
),
IntOrFloat::UnsignedIntType(precision) => Stmt::Let(
assigned,
Expr::Literal(Literal::Int(int)),
Layout::Builtin(int_precision_to_builtin(precision)),
precision.as_layout(),
hole,
),
_ => unreachable!("unexpected float precision for integer"),
@ -2778,7 +2811,7 @@ pub fn with_hole<'a>(
IntOrFloat::BinaryFloatType(precision) => Stmt::Let(
assigned,
Expr::Literal(Literal::Float(float)),
Layout::Builtin(float_precision_to_builtin(precision)),
precision.as_layout(),
hole,
),
IntOrFloat::DecimalFloatType => {
@ -2810,19 +2843,19 @@ pub fn with_hole<'a>(
IntOrFloat::SignedIntType(precision) => Stmt::Let(
assigned,
Expr::Literal(Literal::Int(num.into())),
Layout::Builtin(int_precision_to_builtin(precision)),
precision.as_layout(),
hole,
),
IntOrFloat::UnsignedIntType(precision) => Stmt::Let(
assigned,
Expr::Literal(Literal::Int(num.into())),
Layout::Builtin(int_precision_to_builtin(precision)),
precision.as_layout(),
hole,
),
IntOrFloat::BinaryFloatType(precision) => Stmt::Let(
assigned,
Expr::Literal(Literal::Float(num as f64)),
Layout::Builtin(float_precision_to_builtin(precision)),
precision.as_layout(),
hole,
),
IntOrFloat::DecimalFloatType => {
@ -4131,6 +4164,8 @@ fn construct_closure_data<'a>(
assigned: Symbol,
hole: &'a Stmt<'a>,
) -> Stmt<'a> {
let lambda_set_layout = Layout::LambdaSet(lambda_set);
match lambda_set.layout_for_member(name) {
ClosureRepresentation::Union {
tag_id,
@ -4162,12 +4197,7 @@ fn construct_closure_data<'a>(
arguments: symbols,
};
Stmt::Let(
assigned,
expr,
lambda_set.runtime_representation(),
env.arena.alloc(hole),
)
Stmt::Let(assigned, expr, lambda_set_layout, env.arena.alloc(hole))
}
ClosureRepresentation::AlphabeticOrderStruct(field_layouts) => {
debug_assert_eq!(field_layouts.len(), symbols.len());
@ -4198,7 +4228,7 @@ fn construct_closure_data<'a>(
let expr = Expr::Struct(symbols);
Stmt::Let(assigned, expr, lambda_set.runtime_representation(), hole)
Stmt::Let(assigned, expr, lambda_set_layout, hole)
}
ClosureRepresentation::Other(Layout::Builtin(Builtin::Int1)) => {
debug_assert_eq!(symbols.len(), 0);
@ -4207,7 +4237,7 @@ fn construct_closure_data<'a>(
let tag_id = name != lambda_set.set[0].0;
let expr = Expr::Literal(Literal::Bool(tag_id));
Stmt::Let(assigned, expr, lambda_set.runtime_representation(), hole)
Stmt::Let(assigned, expr, lambda_set_layout, hole)
}
ClosureRepresentation::Other(Layout::Builtin(Builtin::Int8)) => {
debug_assert_eq!(symbols.len(), 0);
@ -4216,7 +4246,7 @@ fn construct_closure_data<'a>(
let tag_id = lambda_set.set.iter().position(|(s, _)| *s == name).unwrap() as u8;
let expr = Expr::Literal(Literal::Byte(tag_id));
Stmt::Let(assigned, expr, lambda_set.runtime_representation(), hole)
Stmt::Let(assigned, expr, lambda_set_layout, hole)
}
_ => unreachable!(),
}
@ -5623,8 +5653,8 @@ fn store_pattern_help<'a>(
// do nothing
return StorePattern::NotProductive(stmt);
}
IntLiteral(_)
| FloatLiteral(_)
IntLiteral(_, _)
| FloatLiteral(_, _)
| DecimalLiteral(_)
| EnumLiteral { .. }
| BitLiteral { .. }
@ -5758,8 +5788,8 @@ fn store_tag_pattern<'a>(
Underscore => {
// ignore
}
IntLiteral(_)
| FloatLiteral(_)
IntLiteral(_, _)
| FloatLiteral(_, _)
| DecimalLiteral(_)
| EnumLiteral { .. }
| BitLiteral { .. }
@ -5834,8 +5864,8 @@ fn store_newtype_pattern<'a>(
Underscore => {
// ignore
}
IntLiteral(_)
| FloatLiteral(_)
IntLiteral(_, _)
| FloatLiteral(_, _)
| DecimalLiteral(_)
| EnumLiteral { .. }
| BitLiteral { .. }
@ -5910,8 +5940,8 @@ fn store_record_destruct<'a>(
// internally. But `y` is never used, so we must make sure it't not stored/loaded.
return StorePattern::NotProductive(stmt);
}
IntLiteral(_)
| FloatLiteral(_)
IntLiteral(_, _)
| FloatLiteral(_, _)
| DecimalLiteral(_)
| EnumLiteral { .. }
| BitLiteral { .. }
@ -6060,7 +6090,7 @@ fn reuse_function_symbol<'a>(
let layout = match raw {
RawFunctionLayout::ZeroArgumentThunk(layout) => layout,
RawFunctionLayout::Function(_, lambda_set, _) => {
lambda_set.runtime_representation()
Layout::LambdaSet(lambda_set)
}
};
@ -6158,7 +6188,7 @@ fn reuse_function_symbol<'a>(
// TODO suspicious
// let layout = Layout::Closure(argument_layouts, lambda_set, ret_layout);
// panic!("suspicious");
let layout = lambda_set.runtime_representation();
let layout = Layout::LambdaSet(lambda_set);
let top_level = ProcLayout::new(env.arena, &[], layout);
procs.insert_passed_by_name(
env,
@ -6178,9 +6208,14 @@ fn reuse_function_symbol<'a>(
layout_cache,
);
// a function name (non-closure) that is passed along
// it never has closure data, so we use the empty struct
return let_empty_struct(symbol, env.arena.alloc(result));
construct_closure_data(
env,
lambda_set,
original,
&[],
symbol,
env.arena.alloc(result),
)
}
}
RawFunctionLayout::ZeroArgumentThunk(ret_layout) => {
@ -6348,7 +6383,7 @@ fn call_by_name<'a>(
procs,
fn_var,
proc_name,
env.arena.alloc(lambda_set.runtime_representation()),
env.arena.alloc(Layout::LambdaSet(lambda_set)),
layout_cache,
assigned,
hole,
@ -6392,7 +6427,7 @@ fn call_by_name<'a>(
procs,
fn_var,
proc_name,
env.arena.alloc(lambda_set.runtime_representation()),
env.arena.alloc(Layout::LambdaSet(lambda_set)),
layout_cache,
closure_data_symbol,
env.arena.alloc(result),
@ -6522,7 +6557,7 @@ fn call_by_name_help<'a>(
force_thunk(
env,
proc_name,
lambda_set.runtime_representation(),
Layout::LambdaSet(lambda_set),
assigned,
hole,
)
@ -6836,13 +6871,7 @@ fn call_specialized_proc<'a>(
arguments: field_symbols,
};
build_call(
env,
call,
assigned,
lambda_set.runtime_representation(),
hole,
)
build_call(env, call, assigned, Layout::LambdaSet(lambda_set), hole)
}
RawFunctionLayout::ZeroArgumentThunk(_) => {
unreachable!()
@ -6882,8 +6911,8 @@ fn call_specialized_proc<'a>(
pub enum Pattern<'a> {
Identifier(Symbol),
Underscore,
IntLiteral(i128),
FloatLiteral(u64),
IntLiteral(i128, IntPrecision),
FloatLiteral(u64, FloatPrecision),
DecimalLiteral(RocDec),
BitLiteral {
value: bool,
@ -6961,22 +6990,36 @@ fn from_can_pattern_help<'a>(
match can_pattern {
Underscore => Ok(Pattern::Underscore),
Identifier(symbol) => Ok(Pattern::Identifier(*symbol)),
IntLiteral(_, _, int) => Ok(Pattern::IntLiteral(*int as i128)),
IntLiteral(var, _, int) => {
match num_argument_to_int_or_float(env.subs, env.ptr_bytes, *var, false) {
IntOrFloat::SignedIntType(precision) | IntOrFloat::UnsignedIntType(precision) => {
Ok(Pattern::IntLiteral(*int as i128, precision))
}
other => {
panic!(
"Invalid precision for int pattern: {:?} has {:?}",
can_pattern, other
)
}
}
}
FloatLiteral(var, float_str, float) => {
// TODO: Can I reuse num_argument_to_int_or_float here if I pass in true?
match num_argument_to_int_or_float(env.subs, env.ptr_bytes, *var, true) {
IntOrFloat::SignedIntType(_) => {
panic!("Invalid percision for float literal = {:?}", var)
IntOrFloat::SignedIntType(_) | IntOrFloat::UnsignedIntType(_) => {
panic!("Invalid precision for float pattern {:?}", var)
}
IntOrFloat::UnsignedIntType(_) => {
panic!("Invalid percision for float literal = {:?}", var)
IntOrFloat::BinaryFloatType(precision) => {
Ok(Pattern::FloatLiteral(f64::to_bits(*float), precision))
}
IntOrFloat::BinaryFloatType(_) => Ok(Pattern::FloatLiteral(f64::to_bits(*float))),
IntOrFloat::DecimalFloatType => {
let dec = match RocDec::from_str(float_str) {
Some(d) => d,
None => panic!("Invalid decimal for float literal = {}. TODO: Make this a nice, user-friendly error message", float_str),
};
Some(d) => d,
None => panic!(
r"Invalid decimal for float literal = {}. TODO: Make this a nice, user-friendly error message",
float_str
),
};
Ok(Pattern::DecimalLiteral(dec))
}
}
@ -6993,9 +7036,15 @@ fn from_can_pattern_help<'a>(
}
NumLiteral(var, num_str, num) => {
match num_argument_to_int_or_float(env.subs, env.ptr_bytes, *var, false) {
IntOrFloat::SignedIntType(_) => Ok(Pattern::IntLiteral(*num as i128)),
IntOrFloat::UnsignedIntType(_) => Ok(Pattern::IntLiteral(*num as i128)),
IntOrFloat::BinaryFloatType(_) => Ok(Pattern::FloatLiteral(*num as u64)),
IntOrFloat::SignedIntType(precision) => {
Ok(Pattern::IntLiteral(*num as i128, precision))
}
IntOrFloat::UnsignedIntType(precision) => {
Ok(Pattern::IntLiteral(*num as i128, precision))
}
IntOrFloat::BinaryFloatType(precision) => {
Ok(Pattern::FloatLiteral(*num as u64, precision))
}
IntOrFloat::DecimalFloatType => {
let dec = match RocDec::from_str(num_str) {
Some(d) => d,
@ -7577,7 +7626,7 @@ fn from_can_record_destruct<'a>(
})
}
#[derive(Debug)]
#[derive(Debug, Clone, Copy, PartialEq, Hash)]
pub enum IntPrecision {
Usize,
I128,
@ -7587,11 +7636,45 @@ pub enum IntPrecision {
I8,
}
impl IntPrecision {
pub fn as_layout(&self) -> Layout<'static> {
Layout::Builtin(self.as_builtin())
}
pub fn as_builtin(&self) -> Builtin<'static> {
use IntPrecision::*;
match self {
I128 => Builtin::Int128,
I64 => Builtin::Int64,
I32 => Builtin::Int32,
I16 => Builtin::Int16,
I8 => Builtin::Int8,
Usize => Builtin::Usize,
}
}
}
#[derive(Debug, Clone, Copy, PartialEq, Hash)]
pub enum FloatPrecision {
F64,
F32,
}
impl FloatPrecision {
pub fn as_layout(&self) -> Layout<'static> {
Layout::Builtin(self.as_builtin())
}
pub fn as_builtin(&self) -> Builtin<'static> {
use FloatPrecision::*;
match self {
F64 => Builtin::Float64,
F32 => Builtin::Float32,
}
}
}
#[derive(Debug)]
pub enum IntOrFloat {
SignedIntType(IntPrecision),
UnsignedIntType(IntPrecision),
@ -7599,26 +7682,6 @@ pub enum IntOrFloat {
DecimalFloatType,
}
fn float_precision_to_builtin(precision: FloatPrecision) -> Builtin<'static> {
use FloatPrecision::*;
match precision {
F64 => Builtin::Float64,
F32 => Builtin::Float32,
}
}
fn int_precision_to_builtin(precision: IntPrecision) -> Builtin<'static> {
use IntPrecision::*;
match precision {
I128 => Builtin::Int128,
I64 => Builtin::Int64,
I32 => Builtin::Int32,
I16 => Builtin::Int16,
I8 => Builtin::Int8,
Usize => Builtin::Usize,
}
}
/// Given the `a` in `Num a`, determines whether it's an int or a float
pub fn num_argument_to_int_or_float(
subs: &Subs,
@ -7915,8 +7978,8 @@ fn match_on_lambda_set<'a>(
let result = union_lambda_set_to_switch(
env,
lambda_set.set,
lambda_set.runtime_representation(),
lambda_set,
Layout::Union(union_layout),
closure_tag_id_symbol,
union_layout.tag_id_layout(),
closure_data_symbol,
@ -7946,6 +8009,7 @@ fn match_on_lambda_set<'a>(
union_lambda_set_branch_help(
env,
function_symbol,
lambda_set,
closure_data_symbol,
Layout::Struct(fields),
argument_symbols,
@ -7994,7 +8058,7 @@ fn match_on_lambda_set<'a>(
#[allow(clippy::too_many_arguments)]
fn union_lambda_set_to_switch<'a>(
env: &mut Env<'a, '_>,
lambda_set: &'a [(Symbol, &'a [Layout<'a>])],
lambda_set: LambdaSet<'a>,
closure_layout: Layout<'a>,
closure_tag_id_symbol: Symbol,
closure_tag_id_layout: Layout<'a>,
@ -8005,7 +8069,7 @@ fn union_lambda_set_to_switch<'a>(
assigned: Symbol,
hole: &'a Stmt<'a>,
) -> Stmt<'a> {
if lambda_set.is_empty() {
if lambda_set.set.is_empty() {
// NOTE this can happen if there is a type error somewhere. Since the lambda set is empty,
// there is really nothing we can do here. We generate a runtime error here which allows
// code gen to proceed. We then assume that we hit another (more descriptive) error before
@ -8017,11 +8081,12 @@ fn union_lambda_set_to_switch<'a>(
let join_point_id = JoinPointId(env.unique_symbol());
let mut branches = Vec::with_capacity_in(lambda_set.len(), env.arena);
let mut branches = Vec::with_capacity_in(lambda_set.set.len(), env.arena);
for (i, (function_symbol, _)) in lambda_set.iter().enumerate() {
for (i, (function_symbol, _)) in lambda_set.set.iter().enumerate() {
let stmt = union_lambda_set_branch(
env,
lambda_set,
join_point_id,
*function_symbol,
closure_data_symbol,
@ -8064,6 +8129,7 @@ fn union_lambda_set_to_switch<'a>(
#[allow(clippy::too_many_arguments)]
fn union_lambda_set_branch<'a>(
env: &mut Env<'a, '_>,
lambda_set: LambdaSet<'a>,
join_point_id: JoinPointId,
function_symbol: Symbol,
closure_data_symbol: Symbol,
@ -8079,6 +8145,7 @@ fn union_lambda_set_branch<'a>(
union_lambda_set_branch_help(
env,
function_symbol,
lambda_set,
closure_data_symbol,
closure_data_layout,
argument_symbols_slice,
@ -8093,6 +8160,7 @@ fn union_lambda_set_branch<'a>(
fn union_lambda_set_branch_help<'a>(
env: &mut Env<'a, '_>,
function_symbol: Symbol,
lambda_set: LambdaSet<'a>,
closure_data_symbol: Symbol,
closure_data_layout: Layout<'a>,
argument_symbols_slice: &'a [Symbol],
@ -8105,12 +8173,19 @@ fn union_lambda_set_branch_help<'a>(
Layout::Struct(&[]) | Layout::Builtin(Builtin::Int1) | Layout::Builtin(Builtin::Int8) => {
(argument_layouts_slice, argument_symbols_slice)
}
_ if lambda_set.member_does_not_need_closure_argument(function_symbol) => {
// sometimes unification causes a function that does not itself capture anything
// to still get a lambda set that does store information. We must not pass a closure
// argument in this case
(argument_layouts_slice, argument_symbols_slice)
}
_ => {
// extend layouts with the layout of the closure environment
let mut argument_layouts =
Vec::with_capacity_in(argument_layouts_slice.len() + 1, env.arena);
argument_layouts.extend(argument_layouts_slice);
argument_layouts.push(closure_data_layout);
argument_layouts.push(Layout::LambdaSet(lambda_set));
// extend symbols with the symbol of the closure environment
let mut argument_symbols =

50
compiler/mono/src/layout.rs
View file

@ -189,6 +189,7 @@ pub enum Layout<'a> {
/// this is important for closures that capture zero-sized values
Struct(&'a [Layout<'a>]),
Union(UnionLayout<'a>),
LambdaSet(LambdaSet<'a>),
RecursivePointer,
}
@ -454,6 +455,17 @@ impl<'a> LambdaSet<'a> {
}
}
pub fn member_does_not_need_closure_argument(&self, function_symbol: Symbol) -> bool {
match self.layout_for_member(function_symbol) {
ClosureRepresentation::Union {
alphabetic_order_fields,
..
} => alphabetic_order_fields.is_empty(),
ClosureRepresentation::AlphabeticOrderStruct(fields) => fields.is_empty(),
ClosureRepresentation::Other(_) => false,
}
}
pub fn layout_for_member(&self, function_symbol: Symbol) -> ClosureRepresentation<'a> {
debug_assert!(
self.set.iter().any(|(s, _)| *s == function_symbol),
@ -531,7 +543,7 @@ impl<'a> LambdaSet<'a> {
_ => {
let mut arguments = Vec::with_capacity_in(argument_layouts.len() + 1, arena);
arguments.extend(argument_layouts);
arguments.push(self.runtime_representation());
arguments.push(Layout::LambdaSet(*self));
arguments.into_bump_slice()
}
@ -587,7 +599,7 @@ impl<'a> LambdaSet<'a> {
// this can happen when there is a type error somewhere
Ok(LambdaSet {
set: &[],
representation: arena.alloc(Layout::Union(UnionLayout::NonRecursive(&[]))),
representation: arena.alloc(Layout::Struct(&[])),
})
}
_ => panic!("called LambdaSet.from_var on invalid input"),
@ -606,7 +618,9 @@ impl<'a> LambdaSet<'a> {
use UnionVariant::*;
match variant {
Never => Layout::Union(UnionLayout::NonRecursive(&[])),
Unit | UnitWithArguments | BoolUnion { .. } | ByteUnion(_) => {
BoolUnion { .. } => Layout::Builtin(Builtin::Int1),
ByteUnion { .. } => Layout::Builtin(Builtin::Int8),
Unit | UnitWithArguments => {
// no useful information to store
Layout::Struct(&[])
}
@ -667,7 +681,6 @@ pub enum Builtin<'a> {
Float128,
Float64,
Float32,
Float16,
Str,
Dict(&'a Layout<'a>, &'a Layout<'a>),
Set(&'a Layout<'a>),
@ -826,6 +839,7 @@ impl<'a> Layout<'a> {
}
}
}
LambdaSet(lambda_set) => lambda_set.runtime_representation().safe_to_memcpy(),
RecursivePointer => {
// We cannot memcpy pointers, because then we would have the same pointer in multiple places!
false
@ -890,6 +904,9 @@ impl<'a> Layout<'a> {
| NonNullableUnwrapped(_) => pointer_size,
}
}
LambdaSet(lambda_set) => lambda_set
.runtime_representation()
.stack_size_without_alignment(pointer_size),
RecursivePointer => pointer_size,
}
}
@ -919,6 +936,9 @@ impl<'a> Layout<'a> {
| NonNullableUnwrapped(_) => pointer_size,
}
}
Layout::LambdaSet(lambda_set) => lambda_set
.runtime_representation()
.alignment_bytes(pointer_size),
Layout::Builtin(builtin) => builtin.alignment_bytes(pointer_size),
Layout::RecursivePointer => pointer_size,
}
@ -929,6 +949,9 @@ impl<'a> Layout<'a> {
Layout::Builtin(builtin) => builtin.allocation_alignment_bytes(pointer_size),
Layout::Struct(_) => unreachable!("not heap-allocated"),
Layout::Union(union_layout) => union_layout.allocation_alignment_bytes(pointer_size),
Layout::LambdaSet(lambda_set) => lambda_set
.runtime_representation()
.allocation_alignment_bytes(pointer_size),
Layout::RecursivePointer => unreachable!("should be looked up to get an actual layout"),
}
}
@ -979,6 +1002,7 @@ impl<'a> Layout<'a> {
| NonNullableUnwrapped(_) => true,
}
}
LambdaSet(lambda_set) => lambda_set.runtime_representation().contains_refcounted(),
RecursivePointer => true,
}
}
@ -1002,6 +1026,7 @@ impl<'a> Layout<'a> {
.append(alloc.text("}"))
}
Union(union_layout) => union_layout.to_doc(alloc, parens),
LambdaSet(lambda_set) => lambda_set.runtime_representation().to_doc(alloc, parens),
RecursivePointer => alloc.text("*self"),
}
}
@ -1093,7 +1118,6 @@ impl<'a> Builtin<'a> {
const F128_SIZE: u32 = 16;
const F64_SIZE: u32 = std::mem::size_of::<f64>() as u32;
const F32_SIZE: u32 = std::mem::size_of::<f32>() as u32;
const F16_SIZE: u32 = 2;
/// Number of machine words in an empty one of these
pub const STR_WORDS: u32 = 2;
@ -1123,7 +1147,6 @@ impl<'a> Builtin<'a> {
Float128 => Builtin::F128_SIZE,
Float64 => Builtin::F64_SIZE,
Float32 => Builtin::F32_SIZE,
Float16 => Builtin::F16_SIZE,
Str | EmptyStr => Builtin::STR_WORDS * pointer_size,
Dict(_, _) | EmptyDict => Builtin::DICT_WORDS * pointer_size,
Set(_) | EmptySet => Builtin::SET_WORDS * pointer_size,
@ -1150,7 +1173,6 @@ impl<'a> Builtin<'a> {
Float128 => align_of::<i128>() as u32,
Float64 => align_of::<f64>() as u32,
Float32 => align_of::<f32>() as u32,
Float16 => align_of::<i16>() as u32,
Dict(_, _) | EmptyDict => pointer_size,
Set(_) | EmptySet => pointer_size,
// we often treat these as i128 (64-bit systems)
@ -1158,8 +1180,8 @@ impl<'a> Builtin<'a> {
//
// In webassembly, For that to be safe
// they must be aligned to allow such access
List(_) | EmptyList => pointer_size.max(8),
Str | EmptyStr => pointer_size.max(8),
List(_) | EmptyList => pointer_size,
Str | EmptyStr => pointer_size,
}
}
@ -1168,7 +1190,7 @@ impl<'a> Builtin<'a> {
match self {
Int128 | Int64 | Int32 | Int16 | Int8 | Int1 | Usize | Decimal | Float128 | Float64
| Float32 | Float16 | EmptyStr | EmptyDict | EmptyList | EmptySet => true,
| Float32 | EmptyStr | EmptyDict | EmptyList | EmptySet => true,
Str | Dict(_, _) | Set(_) | List(_) => false,
}
}
@ -1179,7 +1201,7 @@ impl<'a> Builtin<'a> {
match self {
Int128 | Int64 | Int32 | Int16 | Int8 | Int1 | Usize | Decimal | Float128 | Float64
| Float32 | Float16 | EmptyStr | EmptyDict | EmptyList | EmptySet => false,
| Float32 | EmptyStr | EmptyDict | EmptyList | EmptySet => false,
List(_) => true,
Str | Dict(_, _) | Set(_) => true,
@ -1206,7 +1228,6 @@ impl<'a> Builtin<'a> {
Float128 => alloc.text("Float128"),
Float64 => alloc.text("Float64"),
Float32 => alloc.text("Float32"),
Float16 => alloc.text("Float16"),
EmptyStr => alloc.text("EmptyStr"),
EmptyList => alloc.text("EmptyList"),
@ -1240,8 +1261,7 @@ impl<'a> Builtin<'a> {
| Builtin::Decimal
| Builtin::Float128
| Builtin::Float64
| Builtin::Float32
| Builtin::Float16 => unreachable!("not heap-allocated"),
| Builtin::Float32 => unreachable!("not heap-allocated"),
Builtin::Str => pointer_size,
Builtin::Dict(k, v) => k
.alignment_bytes(pointer_size)
@ -1360,7 +1380,7 @@ fn layout_from_flat_type<'a>(
Func(_, closure_var, _) => {
let lambda_set = LambdaSet::from_var(env.arena, env.subs, closure_var, env.ptr_bytes)?;
Ok(lambda_set.runtime_representation())
Ok(Layout::LambdaSet(lambda_set))
}
Record(fields, ext_var) => {
// extract any values from the ext_var

4
compiler/reporting/src/error/canonicalize.rs
View file

@ -1192,7 +1192,7 @@ fn not_found<'b>(
alloc.reflow(" missing up-top"),
]);
let default_yes = alloc.reflow("these names seem close though:");
let default_yes = alloc.reflow("Did you mean one of these?");
let to_details = |no_suggestion_details, yes_suggestion_details| {
if suggestions.is_empty() {
@ -1240,7 +1240,7 @@ fn module_not_found<'b>(
]);
let default_yes = alloc
.reflow("Is there an import missing? Perhaps there is a typo, these names seem close:");
.reflow("Is there an import missing? Perhaps there is a typo. Did you mean one of these?");
let to_details = |no_suggestion_details, yes_suggestion_details| {
if suggestions.is_empty() {

62
compiler/reporting/src/error/type.rs
View file

@ -1,7 +1,8 @@
use roc_can::expected::{Expected, PExpected};
use roc_collections::all::{Index, MutSet, SendMap};
use roc_module::ident::{IdentStr, Lowercase, TagName};
use roc_module::ident::{Ident, IdentStr, Lowercase, TagName};
use roc_module::symbol::Symbol;
use roc_region::all::{Located, Region};
use roc_solve::solve;
use roc_types::pretty_print::Parens;
use roc_types::types::{Category, ErrorType, PatternCategory, Reason, RecordField, TypeExt};
@ -10,34 +11,39 @@ use std::path::PathBuf;
use crate::report::{Annotation, Report, RocDocAllocator, RocDocBuilder, Severity};
use ven_pretty::DocAllocator;
const DUPLICATE_NAME: &str = "DUPLICATE NAME";
const ADD_ANNOTATIONS: &str = r#"Can more type annotations be added? Type annotations always help me give more specific messages, and I think they could help a lot in this case"#;
pub fn type_problem<'b>(
alloc: &'b RocDocAllocator<'b>,
filename: PathBuf,
problem: solve::TypeError,
) -> Report<'b> {
) -> Option<Report<'b>> {
use solve::TypeError::*;
fn report(title: String, doc: RocDocBuilder<'_>, filename: PathBuf) -> Report<'_> {
Report {
fn report(title: String, doc: RocDocBuilder<'_>, filename: PathBuf) -> Option<Report<'_>> {
Some(Report {
title,
filename,
doc,
severity: Severity::RuntimeError,
}
})
}
match problem {
BadExpr(region, category, found, expected) => {
to_expr_report(alloc, filename, region, category, found, expected)
}
BadPattern(region, category, found, expected) => {
to_pattern_report(alloc, filename, region, category, found, expected)
}
CircularType(region, symbol, overall_type) => {
to_circular_report(alloc, filename, region, symbol, overall_type)
}
BadExpr(region, category, found, expected) => Some(to_expr_report(
alloc, filename, region, category, found, expected,
)),
BadPattern(region, category, found, expected) => Some(to_pattern_report(
alloc, filename, region, category, found, expected,
)),
CircularType(region, symbol, overall_type) => Some(to_circular_report(
alloc,
filename,
region,
symbol,
overall_type,
)),
UnexposedLookup(symbol) => {
let title = "UNRECOGNIZED NAME".to_string();
let doc = alloc
@ -97,12 +103,40 @@ pub fn type_problem<'b>(
report(title, doc, filename)
}
SolvedTypeError => None, // Don't re-report cascading errors - see https://github.com/rtfeldman/roc/pull/1711
Shadowed(original_region, shadow) => {
let doc = report_shadowing(alloc, original_region, shadow);
let title = DUPLICATE_NAME.to_string();
report(title, doc, filename)
}
other => panic!("unhandled bad type: {:?}", other),
}
}
}
}
fn report_shadowing<'b>(
alloc: &'b RocDocAllocator<'b>,
original_region: Region,
shadow: Located<Ident>,
) -> RocDocBuilder<'b> {
let line = r#"Since these types have the same name, it's easy to use the wrong one on accident. Give one of them a new name."#;
alloc.stack(vec![
alloc
.text("The ")
.append(alloc.ident(shadow.value))
.append(alloc.reflow(" name is first defined here:")),
alloc.region(original_region),
alloc.reflow("But then it's defined a second time here:"),
alloc.region(shadow.region),
alloc.reflow(line),
])
}
pub fn cyclic_alias<'b>(
alloc: &'b RocDocAllocator<'b>,
symbol: Symbol,

11
compiler/reporting/src/report.rs
View file

@ -128,6 +128,7 @@ impl<'b> Report<'b> {
pub struct Palette<'a> {
pub primary: &'a str,
pub code_block: &'a str,
pub keyword: &'a str,
pub variable: &'a str,
pub type_variable: &'a str,
pub structure: &'a str,
@ -146,6 +147,7 @@ pub struct Palette<'a> {
pub const DEFAULT_PALETTE: Palette = Palette {
primary: WHITE_CODE,
code_block: WHITE_CODE,
keyword: GREEN_CODE,
variable: BLUE_CODE,
type_variable: YELLOW_CODE,
structure: GREEN_CODE,
@ -810,6 +812,9 @@ where
Symbol => {
self.write_str(self.palette.variable)?;
}
Keyword => {
self.write_str(self.palette.keyword)?;
}
GutterBar => {
self.write_str(self.palette.gutter_bar)?;
}
@ -837,7 +842,7 @@ where
ParserSuggestion => {
self.write_str(self.palette.parser_suggestion)?;
}
TypeBlock | GlobalTag | PrivateTag | RecordField | Keyword => { /* nothing yet */ }
TypeBlock | GlobalTag | PrivateTag | RecordField => { /* nothing yet */ }
}
self.style_stack.push(*annotation);
Ok(())
@ -851,11 +856,11 @@ where
Some(annotation) => match annotation {
Emphasized | Url | TypeVariable | Alias | Symbol | BinOp | Error | GutterBar
| Typo | TypoSuggestion | ParserSuggestion | Structure | CodeBlock | PlainText
| LineNumber | Tip | Module | Header => {
| LineNumber | Tip | Module | Header | Keyword => {
self.write_str(RESET_CODE)?;
}
TypeBlock | GlobalTag | PrivateTag | RecordField | Keyword => { /* nothing yet */ }
TypeBlock | GlobalTag | PrivateTag | RecordField => { /* nothing yet */ }
},
}
Ok(())

19
compiler/reporting/tests/test_reporting.rs
View file

@ -154,8 +154,9 @@ mod test_reporting {
}
for problem in type_problems {
let report = type_problem(&alloc, filename.clone(), problem.clone());
reports.push(report);
if let Some(report) = type_problem(&alloc, filename.clone(), problem.clone()) {
reports.push(report);
}
}
for problem in mono_problems {
@ -541,7 +542,7 @@ mod test_reporting {
8│ 4 -> bar baz "yay"
^^^
these names seem close though:
Did you mean one of these?
baz
Nat
@ -739,7 +740,7 @@ mod test_reporting {
<cyan>3<reset><cyan>│<reset> <white>theAdmin<reset>
<red>^^^^^^^^<reset>
these names seem close though:
Did you mean one of these?
Decimal
Dec
@ -1491,7 +1492,7 @@ mod test_reporting {
2│ { foo: 2 } -> foo
^^^
these names seem close though:
Did you mean one of these?
Bool
U8
@ -1947,7 +1948,7 @@ mod test_reporting {
2│ f = \_ -> ok 4
^^
these names seem close though:
Did you mean one of these?
U8
f
@ -3634,8 +3635,8 @@ mod test_reporting {
1│ Foo.test
^^^^^^^^
Is there an import missing? Perhaps there is a typo, these names seem
close:
Is there an import missing? Perhaps there is a typo. Did you mean one
of these?
Bool
Num
@ -5797,7 +5798,7 @@ mod test_reporting {
1│ [ "foo", bar("") ]
^^^
these names seem close though:
Did you mean one of these?
Nat
Str

42
compiler/test_gen/src/gen_list.rs
View file

@ -2017,3 +2017,45 @@ fn lists_with_incompatible_type_param_in_if() {
RocStr
);
}
#[test]
fn map_with_index_multi_record() {
// see https://github.com/rtfeldman/roc/issues/1700
assert_evals_to!(
indoc!(
r#"
List.mapWithIndex [ { x: {}, y: {} } ] \_, _ -> {}
"#
),
RocList::from_slice(&[((), ())]),
RocList<((), ())>
);
}
#[test]
fn empty_list_of_function_type() {
// see https://github.com/rtfeldman/roc/issues/1732
assert_evals_to!(
indoc!(
r#"
myList : List (Str -> Str)
myList = []
myClosure : Str -> Str
myClosure = \_ -> "bar"
choose =
if False then
myList
else
[ myClosure ]
when List.get choose 0 is
Ok f -> f "foo"
Err _ -> "bad!"
"#
),
RocStr::from_slice(b"bar"),
RocStr
);
}

44
compiler/test_gen/src/gen_num.rs
View file

@ -1778,4 +1778,48 @@ mod gen_num {
u32
);
}
#[test]
fn when_on_i32() {
assert_evals_to!(
indoc!(
r#"
app "test" provides [ main ] to "./platform"
x : I32
x = 0
main : I32
main =
when x is
0 -> 42
_ -> -1
"#
),
42,
i32
);
}
#[test]
fn when_on_i16() {
assert_evals_to!(
indoc!(
r#"
app "test" provides [ main ] to "./platform"
x : I16
x = 0
main : I16
main =
when x is
0 -> 42
_ -> -1
"#
),
42,
i16
);
}
}

129
compiler/test_gen/src/gen_primitives.rs
View file

@ -2531,6 +2531,8 @@ fn pattern_match_unit_tag() {
);
}
// see for why this is disabled on wasm32 https://github.com/rtfeldman/roc/issues/1687
#[cfg(not(feature = "wasm-cli-run"))]
#[test]
fn mirror_llvm_alignment_padding() {
// see https://github.com/rtfeldman/roc/issues/1569
@ -2616,7 +2618,8 @@ fn lambda_set_struct_byte() {
r = Red
p1 = (\u -> r == u)
oneOfResult = List.map [p1, p1] (\p -> p Green)
foobarbaz = (\p -> p Green)
oneOfResult = List.map [p1, p1] foobarbaz
when oneOfResult is
_ -> 32
@ -2779,3 +2782,127 @@ fn value_not_exposed_hits_panic() {
i64
);
}
#[test]
fn mix_function_and_closure() {
// see https://github.com/rtfeldman/roc/pull/1706
assert_evals_to!(
indoc!(
r#"
app "test" provides [ main ] to "./platform"
# foo does not capture any variables
# but through unification will get a lambda set that does store information
# we must handle that correctly
foo = \x -> x
bar = \y -> \_ -> y
main : Str
main =
(if 1 == 1 then foo else (bar "nope nope nope")) "hello world"
"#
),
RocStr::from_slice(b"hello world"),
RocStr
);
}
#[test]
fn mix_function_and_closure_level_of_indirection() {
// see https://github.com/rtfeldman/roc/pull/1706
assert_evals_to!(
indoc!(
r#"
app "test" provides [ main ] to "./platform"
foo = \x -> x
bar = \y -> \_ -> y
f = (if 1 == 1 then foo else (bar "nope nope nope"))
main : Str
main =
f "hello world"
"#
),
RocStr::from_slice(b"hello world"),
RocStr
);
}
#[test]
fn do_pass_bool_byte_closure_layout() {
// see https://github.com/rtfeldman/roc/pull/1706
// the distinction is actually important, dropping that info means some functions just get
// skipped
assert_evals_to!(
indoc!(
r#"
app "test" provides [ main ] to "./platform"
## PARSER
Parser a : List U8 -> List [Pair a (List U8)]
## ANY
# If succcessful, the any parser consumes one character
any: Parser U8
any = \inp ->
when List.first inp is
Ok u -> [Pair u (List.drop inp 1)]
_ -> [ ]
## SATISFY
satisfy : (U8 -> Bool) -> Parser U8
satisfy = \predicate ->
\input ->
walker = \(Pair u rest), accum ->
if predicate u then
Stop [ Pair u rest ]
else
Stop accum
List.walkUntil (any input) walker []
oneOf : List (Parser a) -> Parser a
oneOf = \parserList ->
\input ->
walker = \p, accum ->
output = p input
if List.len output == 1 then
Stop output
else
Continue accum
List.walkUntil parserList walker []
satisfyA = satisfy (\u -> u == 97) # recognize 97
satisfyB = satisfy (\u -> u == 98) # recognize 98
test1 = if List.len ((oneOf [satisfyA, satisfyB]) [97, 98, 99, 100] ) == 1 then "PASS" else "FAIL"
test2 = if List.len ((oneOf [satisfyA, satisfyB]) [98, 99, 100, 97] ) == 1 then "PASS" else "FAIL"
test3 = if List.len ((oneOf [satisfyB , satisfyA]) [98, 99, 100, 97] ) == 1 then "PASS" else "FAIL"
test4 = if List.len ((oneOf [satisfyA, satisfyB]) [99, 100, 101] ) == 0 then "PASS" else "FAIL"
main : Str
main = [test1, test2, test3, test4] |> Str.joinWith ", "
"#
),
RocStr::from_slice(b"PASS, PASS, PASS, PASS"),
RocStr
);
}

28
compiler/test_gen/src/gen_str.rs
View file

@ -949,3 +949,31 @@ fn str_from_utf8_range_count_too_high_for_start() {
RocStr
);
}
#[test]
fn str_repeat_small() {
assert_evals_to!(
indoc!(r#"Str.repeat "Roc" 3"#),
RocStr::from("RocRocRoc"),
RocStr
);
}
#[test]
fn str_repeat_big() {
assert_evals_to!(
indoc!(r#"Str.repeat "more than 16 characters" 2"#),
RocStr::from("more than 16 charactersmore than 16 characters"),
RocStr
);
}
#[test]
fn str_repeat_empty_string() {
assert_evals_to!(indoc!(r#"Str.repeat "" 3"#), RocStr::from(""), RocStr);
}
#[test]
fn str_repeat_zero_times() {
assert_evals_to!(indoc!(r#"Str.repeat "Roc" 0"#), RocStr::from(""), RocStr);
}

9
compiler/test_gen/src/helpers/eval.rs
View file

@ -143,12 +143,13 @@ fn create_llvm_module<'a>(
}
for problem in type_problems {
let report = type_problem(&alloc, module_path.clone(), problem);
let mut buf = String::new();
if let Some(report) = type_problem(&alloc, module_path.clone(), problem) {
let mut buf = String::new();
report.render_color_terminal(&mut buf, &alloc, &palette);
report.render_color_terminal(&mut buf, &alloc, &palette);
lines.push(buf);
lines.push(buf);
}
}
for problem in mono_problems {

43
compiler/test_mono/generated/empty_list_of_function_type.txt Normal file
View file

@ -0,0 +1,43 @@
procedure List.3 (#Attr.2, #Attr.3):
let Test.20 = lowlevel ListLen #Attr.2;
let Test.17 = lowlevel NumLt #Attr.3 Test.20;
if Test.17 then
let Test.19 = lowlevel ListGetUnsafe #Attr.2 #Attr.3;
let Test.18 = Ok Test.19;
ret Test.18;
else
let Test.16 = Struct {};
let Test.15 = Err Test.16;
ret Test.15;
procedure Test.2 (Test.6):
let Test.24 = "bar";
ret Test.24;
procedure Test.0 ():
let Test.1 = Array [];
joinpoint Test.22 Test.3:
let Test.14 = 0i64;
let Test.7 = CallByName List.3 Test.3 Test.14;
dec Test.3;
let Test.11 = 1i64;
let Test.12 = GetTagId Test.7;
let Test.13 = lowlevel Eq Test.11 Test.12;
if Test.13 then
let Test.5 = UnionAtIndex (Id 1) (Index 0) Test.7;
let Test.9 = "foo";
let Test.8 = CallByName Test.2 Test.9;
dec Test.9;
ret Test.8;
else
let Test.10 = "bad!";
ret Test.10;
in
let Test.25 = false;
if Test.25 then
jump Test.22 Test.1;
else
dec Test.1;
let Test.23 = Struct {};
let Test.21 = Array [Test.23];
jump Test.22 Test.21;

4
compiler/test_mono/generated/quicksort_help.txt
View file

@ -10,7 +10,7 @@ procedure Num.27 (#Attr.2, #Attr.3):
let Test.26 = lowlevel NumLt #Attr.2 #Attr.3;
ret Test.26;
procedure Test.1 (Test.29, Test.30, Test.31):
procedure Test.1 (Test.27, Test.28, Test.29):
joinpoint Test.12 Test.2 Test.3 Test.4:
let Test.14 = CallByName Num.27 Test.3 Test.4;
if Test.14 then
@ -29,7 +29,7 @@ procedure Test.1 (Test.29, Test.30, Test.31):
else
ret Test.2;
in
jump Test.12 Test.29 Test.30 Test.31;
jump Test.12 Test.27 Test.28 Test.29;
procedure Test.0 ():
let Test.9 = Array [];

66
compiler/test_mono/generated/somehow_drops_definitions.txt
View file

@ -1,43 +1,53 @@
procedure Num.24 (#Attr.2, #Attr.3):
let Test.24 = lowlevel NumAdd #Attr.2 #Attr.3;
ret Test.24;
let Test.27 = lowlevel NumAdd #Attr.2 #Attr.3;
ret Test.27;
procedure Num.26 (#Attr.2, #Attr.3):
let Test.19 = lowlevel NumMul #Attr.2 #Attr.3;
ret Test.19;
procedure Test.1 ():
let Test.25 = 1i64;
ret Test.25;
procedure Test.2 ():
let Test.20 = 2i64;
ret Test.20;
procedure Test.3 (Test.6):
let Test.23 = CallByName Test.1;
let Test.22 = CallByName Num.24 Test.6 Test.23;
let Test.22 = lowlevel NumMul #Attr.2 #Attr.3;
ret Test.22;
procedure Test.1 ():
let Test.28 = 1i64;
ret Test.28;
procedure Test.2 ():
let Test.23 = 2i64;
ret Test.23;
procedure Test.3 (Test.6):
let Test.26 = CallByName Test.1;
let Test.25 = CallByName Num.24 Test.6 Test.26;
ret Test.25;
procedure Test.4 (Test.7):
let Test.18 = CallByName Test.2;
let Test.17 = CallByName Num.26 Test.7 Test.18;
ret Test.17;
let Test.21 = CallByName Test.2;
let Test.20 = CallByName Num.26 Test.7 Test.21;
ret Test.20;
procedure Test.5 (Test.8, Test.9):
let Test.14 = CallByName Test.3 Test.9;
ret Test.14;
joinpoint Test.15 Test.14:
ret Test.14;
in
switch Test.8:
case 0:
let Test.16 = CallByName Test.3 Test.9;
jump Test.15 Test.16;
default:
let Test.17 = CallByName Test.4 Test.9;
jump Test.15 Test.17;
procedure Test.0 ():
joinpoint Test.16 Test.12:
joinpoint Test.19 Test.12:
let Test.13 = 42i64;
let Test.11 = CallByName Test.5 Test.12 Test.13;
ret Test.11;
in
let Test.21 = true;
if Test.21 then
let Test.3 = Struct {};
jump Test.16 Test.3;
let Test.24 = true;
if Test.24 then
let Test.3 = false;
jump Test.19 Test.3;
else
let Test.4 = Struct {};
jump Test.16 Test.4;
let Test.4 = true;
jump Test.19 Test.4;

27
compiler/test_mono/src/lib.rs
View file

@ -1084,6 +1084,33 @@ fn specialize_lowlevel() {
)
}
#[mono_test]
fn empty_list_of_function_type() {
// see https://github.com/rtfeldman/roc/issues/1732
indoc!(
r#"
app "test" provides [ main ] to "./platform"
main =
myList : List (Str -> Str)
myList = []
myClosure : Str -> Str
myClosure = \_ -> "bar"
choose =
if False then
myList
else
[ myClosure ]
when List.get choose 0 is
Ok f -> f "foo"
Err _ -> "bad!"
"#
)
}
// #[ignore]
// #[mono_test]
// fn static_str_closure() {

4
examples/.gitignore vendored
View file

@ -4,3 +4,7 @@ app
libhost.a
roc_app.ll
roc_app.bc
dynhost
preprocessedhost
metadata
libapp.so

45
examples/benchmarks/platform/host.zig
View file

@ -23,16 +23,19 @@ comptime {
const mem = std.mem;
const Allocator = mem.Allocator;
extern fn roc__mainForHost_1_exposed([*]u8) void;
extern fn roc__mainForHost_1_exposed_generic([*]u8) void;
extern fn roc__mainForHost_size() i64;
extern fn roc__mainForHost_1_Fx_caller(*const u8, [*]u8, [*]u8) void;
extern fn roc__mainForHost_1_Fx_size() i64;
extern fn roc__mainForHost_1_Fx_result_size() i64;
const Align = extern struct { a: usize, b: usize };
extern fn malloc(size: usize) callconv(.C) ?*align(@alignOf(Align)) c_void;
extern fn realloc(c_ptr: [*]align(@alignOf(Align)) u8, size: usize) callconv(.C) ?*c_void;
extern fn free(c_ptr: [*]align(@alignOf(Align)) u8) callconv(.C) void;
const Align = 2 * @alignOf(usize);
extern fn malloc(size: usize) callconv(.C) ?*align(Align) c_void;
extern fn realloc(c_ptr: [*]align(Align) u8, size: usize) callconv(.C) ?*c_void;
extern fn free(c_ptr: [*]align(Align) u8) callconv(.C) void;
extern fn memcpy(dst: [*]u8, src: [*]u8, size: usize) callconv(.C) void;
extern fn memset(dst: [*]u8, value: i32, size: usize) callconv(.C) void;
const DEBUG: bool = false;
@ -53,7 +56,7 @@ export fn roc_realloc(c_ptr: *c_void, new_size: usize, old_size: usize, alignmen
stdout.print("realloc: {d} (alignment {d}, old_size {d})\n", .{ c_ptr, alignment, old_size }) catch unreachable;
}
return realloc(@alignCast(@alignOf(Align), @ptrCast([*]u8, c_ptr)), new_size);
return realloc(@alignCast(Align, @ptrCast([*]u8, c_ptr)), new_size);
}
export fn roc_dealloc(c_ptr: *c_void, alignment: u32) callconv(.C) void {
@ -62,7 +65,7 @@ export fn roc_dealloc(c_ptr: *c_void, alignment: u32) callconv(.C) void {
stdout.print("dealloc: {d} (alignment {d})\n", .{ c_ptr, alignment }) catch unreachable;
}
free(@alignCast(@alignOf(Align), @ptrCast([*]u8, c_ptr)));
free(@alignCast(Align, @ptrCast([*]u8, c_ptr)));
}
export fn roc_panic(c_ptr: *c_void, tag_id: u32) callconv(.C) void {
@ -74,9 +77,17 @@ export fn roc_panic(c_ptr: *c_void, tag_id: u32) callconv(.C) void {
std.process.exit(0);
}
export fn roc_memcpy(dst: [*]u8, src: [*]u8, size: usize) callconv(.C) void{
return memcpy(dst, src, size);
}
export fn roc_memset(dst: [*]u8, value: i32, size: usize) callconv(.C) void{
return memset(dst, value, size);
}
const Unit = extern struct {};
pub fn main() u8 {
pub export fn main() callconv(.C) u8 {
const allocator = std.heap.page_allocator;
const size = @intCast(usize, roc__mainForHost_size());
@ -90,23 +101,11 @@ pub fn main() u8 {
var ts1: std.os.timespec = undefined;
std.os.clock_gettime(std.os.CLOCK_REALTIME, &ts1) catch unreachable;
roc__mainForHost_1_exposed(output);
roc__mainForHost_1_exposed_generic(output);
const flag = @ptrCast(*u64, @alignCast(@alignOf(u64), output)).*;
const closure_data_pointer = @ptrCast([*]u8, output);
if (flag == 0) {
// all is well
const closure_data_pointer = @ptrCast([*]u8, output[@sizeOf(u64)..size]);
call_the_closure(closure_data_pointer);
} else {
const ptr = @ptrCast(*u32, output + @sizeOf(u64));
const msg = @intToPtr([*:0]const u8, ptr.*);
const stderr = std.io.getStdErr().writer();
stderr.print("Application crashed with message\n\n {s}\n\nShutting down\n", .{msg}) catch unreachable;
return 0;
}
call_the_closure(closure_data_pointer);
var ts2: std.os.timespec = undefined;
std.os.clock_gettime(std.os.CLOCK_REALTIME, &ts2) catch unreachable;

30
examples/cli/platform/src/lib.rs
View file

@ -4,7 +4,7 @@ use core::alloc::Layout;
use core::ffi::c_void;
use core::mem::MaybeUninit;
use libc;
use roc_std::{RocCallResult, RocStr};
use roc_std::RocStr;
use std::ffi::CStr;
use std::os::raw::c_char;
@ -70,23 +70,11 @@ pub fn rust_main() -> isize {
roc_main(buffer);
let output = buffer as *mut RocCallResult<()>;
let result = call_the_closure(buffer);
match (&*output).into() {
Ok(()) => {
let closure_data_ptr = buffer.offset(8);
let result = call_the_closure(closure_data_ptr as *const u8);
std::alloc::dealloc(buffer, layout);
std::alloc::dealloc(buffer, layout);
result
}
Err(msg) => {
std::alloc::dealloc(buffer, layout);
panic!("Roc failed with message: {}", msg);
}
}
result
};
// Exit code
@ -105,15 +93,9 @@ unsafe fn call_the_closure(closure_data_ptr: *const u8) -> i64 {
buffer as *mut u8,
);
let output = &*(buffer as *mut RocCallResult<()>);
std::alloc::dealloc(buffer, layout);
match output.into() {
Ok(_) => {
std::alloc::dealloc(buffer, layout);
0
}
Err(e) => panic!("failed with {}", e),
}
0
}
#[no_mangle]

5
examples/effect/Main.roc
View file

@ -5,4 +5,7 @@ app "effect-example"
main : Effect.Effect {}
main =
Effect.after Effect.getLine \lineThisThing -> Effect.putLine lineThisThing
Effect.after (Effect.getLine) \line ->
Effect.after (Effect.putLine "You entered: \(line)") \{} ->
Effect.after (Effect.putLine "It is known") \{} ->
Effect.always {}

52
examples/effect/thing/platform-dir/host.zig
View file

@ -32,6 +32,8 @@ extern fn roc__mainForHost_1_Fx_result_size() i64;
extern fn malloc(size: usize) callconv(.C) ?*c_void;
extern fn realloc(c_ptr: [*]align(@alignOf(u128)) u8, size: usize) callconv(.C) ?*c_void;
extern fn free(c_ptr: [*]align(@alignOf(u128)) u8) callconv(.C) void;
extern fn memcpy(dst: [*]u8, src: [*]u8, size: usize) callconv(.C) void;
extern fn memset(dst: [*]u8, value: i32, size: usize) callconv(.C) void;
export fn roc_alloc(size: usize, alignment: u32) callconv(.C) ?*c_void {
return malloc(size);
@ -52,18 +54,29 @@ export fn roc_panic(c_ptr: *c_void, tag_id: u32) callconv(.C) void {
std.process.exit(0);
}
export fn roc_memcpy(dst: [*]u8, src: [*]u8, size: usize) callconv(.C) void{
return memcpy(dst, src, size);
}
export fn roc_memset(dst: [*]u8, value: i32, size: usize) callconv(.C) void{
return memset(dst, value, size);
}
const Unit = extern struct {};
pub export fn main() u8 {
const allocator = std.heap.page_allocator;
const stdout = std.io.getStdOut().writer();
const stderr = std.io.getStdErr().writer();
const size = @intCast(usize, roc__mainForHost_size());
const raw_output = std.heap.c_allocator.alloc(u8, size) catch unreachable;
// NOTE the return size can be zero, which will segfault. Always allocate at least 8 bytes
const size = std.math.max(8, @intCast(usize, roc__mainForHost_size()));
const raw_output = allocator.allocAdvanced(u8, @alignOf(u64), @intCast(usize, size), .at_least) catch unreachable;
var output = @ptrCast([*]u8, raw_output);
defer {
std.heap.c_allocator.free(raw_output);
allocator.free(raw_output);
}
var ts1: std.os.timespec = undefined;
@ -71,21 +84,7 @@ pub export fn main() u8 {
roc__mainForHost_1_exposed(output);
const elements = @ptrCast([*]u64, @alignCast(8, output));
var flag = elements[0];
if (flag == 0) {
// all is well
const closure_data_pointer = @ptrCast([*]u8, output[8..size]);
call_the_closure(closure_data_pointer);
} else {
const msg = @intToPtr([*:0]const u8, elements[1]);
stderr.print("Application crashed with message\n\n {s}\n\nShutting down\n", .{msg}) catch unreachable;
return 0;
}
call_the_closure(output);
var ts2: std.os.timespec = undefined;
std.os.clock_gettime(std.os.CLOCK_REALTIME, &ts2) catch unreachable;
@ -102,27 +101,20 @@ fn to_seconds(tms: std.os.timespec) f64 {
}
fn call_the_closure(closure_data_pointer: [*]u8) void {
const allocator = std.heap.page_allocator;
const size = roc__mainForHost_1_Fx_result_size();
const raw_output = std.heap.c_allocator.alloc(u8, @intCast(usize, size)) catch unreachable;
const raw_output = allocator.allocAdvanced(u8, @alignOf(u64), @intCast(usize, size), .at_least) catch unreachable;
var output = @ptrCast([*]u8, raw_output);
defer {
std.heap.c_allocator.free(raw_output);
allocator.free(raw_output);
}
const flags: u8 = 0;
roc__mainForHost_1_Fx_caller(&flags, closure_data_pointer, output);
const elements = @ptrCast([*]u64, @alignCast(8, output));
var flag = elements[0];
if (flag == 0) {
return;
} else {
unreachable;
}
return;
}
pub export fn roc_fx_getLine() str.RocStr {

2
examples/fib/.gitignore vendored Normal file
View file

@ -0,0 +1,2 @@
add
fib

15
examples/fib/Fib.roc Normal file
View file

@ -0,0 +1,15 @@
app "fib"
packages { base: "platform" }
imports []
provides [ main ] to base
main = \n -> fib n 0 1
# the clever implementation requires join points
fib = \n, a, b ->
if n == 0 then
a
else
fib (n - 1) b (a + b)

10
examples/fib/platform/Package-Config.roc Normal file
View file

@ -0,0 +1,10 @@
platform examples/add
requires {}{ main : I64 -> I64 }
exposes []
packages {}
imports []
provides [ mainForHost ]
effects fx.Effect {}
mainForHost : I64 -> I64
mainForHost = \a -> main a

97
examples/fib/platform/host.zig Normal file
View file

@ -0,0 +1,97 @@
const std = @import("std");
const testing = std.testing;
const expectEqual = testing.expectEqual;
const expect = testing.expect;
comptime {
// This is a workaround for https://github.com/ziglang/zig/issues/8218
// which is only necessary on macOS.
//
// Once that issue is fixed, we can undo the changes in
// 177cf12e0555147faa4d436e52fc15175c2c4ff0 and go back to passing
// -fcompiler-rt in link.rs instead of doing this. Note that this
// workaround is present in many host.zig files, so make sure to undo
// it everywhere!
if (std.builtin.os.tag == .macos) {
_ = @import("compiler_rt");
}
}
const mem = std.mem;
const Allocator = mem.Allocator;
// NOTE the LLVM backend expects this signature
// extern fn roc__mainForHost_1_exposed(i64, *i64) void;
extern fn roc__mainForHost_1_exposed(i64) i64;
const Align = extern struct { a: usize, b: usize };
extern fn malloc(size: usize) callconv(.C) ?*align(@alignOf(Align)) c_void;
extern fn realloc(c_ptr: [*]align(@alignOf(Align)) u8, size: usize) callconv(.C) ?*c_void;
extern fn free(c_ptr: [*]align(@alignOf(Align)) u8) callconv(.C) void;
const DEBUG: bool = false;
export fn roc_alloc(size: usize, alignment: u32) callconv(.C) ?*c_void {
if (DEBUG) {
var ptr = malloc(size);
const stdout = std.io.getStdOut().writer();
stdout.print("alloc: {d} (alignment {d}, size {d})\n", .{ ptr, alignment, size }) catch unreachable;
return ptr;
} else {
return malloc(size);
}
}
export fn roc_realloc(c_ptr: *c_void, new_size: usize, old_size: usize, alignment: u32) callconv(.C) ?*c_void {
if (DEBUG) {
const stdout = std.io.getStdOut().writer();
stdout.print("realloc: {d} (alignment {d}, old_size {d})\n", .{ c_ptr, alignment, old_size }) catch unreachable;
}
return realloc(@alignCast(@alignOf(Align), @ptrCast([*]u8, c_ptr)), new_size);
}
export fn roc_dealloc(c_ptr: *c_void, alignment: u32) callconv(.C) void {
if (DEBUG) {
const stdout = std.io.getStdOut().writer();
stdout.print("dealloc: {d} (alignment {d})\n", .{ c_ptr, alignment }) catch unreachable;
}
free(@alignCast(@alignOf(Align), @ptrCast([*]u8, c_ptr)));
}
export fn roc_panic(c_ptr: *c_void, tag_id: u32) callconv(.C) void {
_ = tag_id;
const stderr = std.io.getStdErr().writer();
const msg = @ptrCast([*:0]const u8, c_ptr);
stderr.print("Application crashed with message\n\n {s}\n\nShutting down\n", .{msg}) catch unreachable;
std.process.exit(0);
}
pub export fn main() u8 {
const stdout = std.io.getStdOut().writer();
const stderr = std.io.getStdErr().writer();
// start time
var ts1: std.os.timespec = undefined;
std.os.clock_gettime(std.os.CLOCK_REALTIME, &ts1) catch unreachable;
const result = roc__mainForHost_1_exposed(10);
stdout.print("{d}\n", .{result}) catch unreachable;
// end time
var ts2: std.os.timespec = undefined;
std.os.clock_gettime(std.os.CLOCK_REALTIME, &ts2) catch unreachable;
const delta = to_seconds(ts2) - to_seconds(ts1);
stderr.print("runtime: {d:.3}ms\n", .{delta * 1000}) catch unreachable;
return 0;
}
fn to_seconds(tms: std.os.timespec) f64 {
return @intToFloat(f64, tms.tv_sec) + (@intToFloat(f64, tms.tv_nsec) / 1_000_000_000.0);
}

9
examples/hello-rust/platform/host.c
View file

@ -1,7 +1,12 @@
#include <stdio.h>
#include <string.h>
extern int rust_main();
int main() {
return rust_main();
int main() { return rust_main(); }
void *roc_memcpy(void *dest, const void *src, size_t n) {
return memcpy(dest, src, n);
}
void *roc_memset(void *str, int c, size_t n) { return memset(str, c, n); }

25
examples/hello-rust/platform/src/lib.rs
View file

@ -3,12 +3,12 @@
use core::ffi::c_void;
use core::mem::MaybeUninit;
use libc::c_char;
use roc_std::{RocCallResult, RocStr};
use roc_std::RocStr;
use std::ffi::CStr;
extern "C" {
#[link_name = "roc__mainForHost_1_exposed"]
fn roc_main(output: *mut RocCallResult<RocStr>) -> ();
fn roc_main() -> RocStr;
}
#[no_mangle]
@ -46,25 +46,14 @@ pub unsafe fn roc_panic(c_ptr: *mut c_void, tag_id: u32) {
#[no_mangle]
pub fn rust_main() -> isize {
let mut call_result: MaybeUninit<RocCallResult<RocStr>> = MaybeUninit::uninit();
unsafe {
roc_main(call_result.as_mut_ptr());
let roc_str = roc_main();
let output = call_result.assume_init();
let len = roc_str.len();
let str_bytes = roc_str.get_bytes() as *const libc::c_void;
match output.into() {
Ok(roc_str) => {
let len = roc_str.len();
let str_bytes = roc_str.get_bytes() as *const libc::c_void;
if libc::write(1, str_bytes, len) < 0 {
panic!("Writing to stdout failed!");
}
}
Err(msg) => {
panic!("Roc failed with message: {}", msg);
}
if libc::write(1, str_bytes, len) < 0 {
panic!("Writing to stdout failed!");
}
}

2
examples/hello-web/.gitignore vendored Normal file
View file

@ -0,0 +1,2 @@
hello-web
*.wat

12
examples/hello-web/Hello.roc Normal file
View file

@ -0,0 +1,12 @@
app "hello-web"
packages { base: "platform" }
imports []
provides [ main ] to base
greeting =
hi = "Hello"
name = "World"
"\(hi), \(name)!"
main = greeting

49
examples/hello-web/README.md Normal file
View file

@ -0,0 +1,49 @@
# Hello, World!
To run, go to the project home directory and run:
```bash
$ cargo run -- build --backend=wasm32 examples/hello-web/Hello.roc
```
Then `cd` into the example directory and run any web server that can handle WebAssembly.
For example with `http-server`:
```bash
cd examples/hello-web
npm install -g http-server
http-server
```
Now open your browser at http://localhost:8080
## Design Notes
This demonstrates the basic design of hosts: Roc code gets compiled into a pure
function (in this case, a thunk that always returns `"Hello, World!"`) and
then the host calls that function. Fundamentally, that's the whole idea! The host
might not even have a `main` - it could be a library, a plugin, anything.
Everything else is built on this basic "hosts calling linked pure functions" design.
For example, things get more interesting when the compiled Roc function returns
a `Task` - that is, a tagged union data structure containing function pointers
to callback closures. This lets the Roc pure function describe arbitrary
chainable effects, which the host can interpret to perform I/O as requested by
the Roc program. (The tagged union `Task` would have a variant for each supported
I/O operation.)
In this trivial example, it's very easy to line up the API between the host and
the Roc program. In a more involved host, this would be much trickier - especially
if the API were changing frequently during development.
The idea there is to have a first-class concept of "glue code" which host authors
can write (it would be plain Roc code, but with some extra keywords that aren't
available in normal modules - kinda like `port module` in Elm), and which
describe both the Roc-host/C boundary as well as the Roc-host/Roc-app boundary.
Roc application authors only care about the Roc-host/Roc-app portion, and the
host author only cares about the Roc-host/C boundary when implementing the host.
Using this glue code, the Roc compiler can generate C header files describing the
boundary. This not only gets us host compatibility with C compilers, but also
Rust FFI for free, because [`rust-bindgen`](https://github.com/rust-lang/rust-bindgen)
generates correct Rust FFI bindings from C headers.

12
examples/hello-web/index.html Normal file
View file

@ -0,0 +1,12 @@
<html>
<body>
<div id="output"></div>
<script src="platform/host.js"></script>
<script>
const elem = document.getElementById("output");
roc_web_platform_run("./hello-web.wasm", (string_from_roc) => {
elem.textContent = string_from_roc;
});
</script>
</body>
</html>

10
examples/hello-web/platform/Package-Config.roc Normal file
View file

@ -0,0 +1,10 @@
platform examples/hello-world
requires {}{ main : Str }
exposes []
packages {}
imports []
provides [ mainForHost ]
effects fx.Effect {}
mainForHost : Str
mainForHost = main

57
examples/hello-web/platform/host.js Normal file
View file

@ -0,0 +1,57 @@
async function roc_web_platform_run(wasm_filename, callback) {
const decoder = new TextDecoder();
let memory_bytes;
let exit_code;
function js_display_roc_string(str_bytes, str_len) {
const utf8_bytes = memory_bytes.subarray(str_bytes, str_bytes + str_len);
const js_string = decoder.decode(utf8_bytes);
callback(js_string);
}
const importObj = {
wasi_snapshot_preview1: {
proc_exit: (code) => {
if (code !== 0) {
console.error(`Exited with code ${code}`);
}
exit_code = code;
},
roc_panic: (_pointer, _tag_id) => {
throw 'Roc panicked!';
}
},
env: {
js_display_roc_string,
},
};
let wasm;
const response = await fetch(wasm_filename);
if (WebAssembly.instantiateStreaming) {
// streaming API has better performance if available
wasm = await WebAssembly.instantiateStreaming(response, importObj);
} else {
const module_bytes = await response.arrayBuffer();
wasm = await WebAssembly.instantiate(module_bytes, importObj);
}
memory_bytes = new Uint8Array(wasm.instance.exports.memory.buffer);
try {
wasm.instance.exports._start();
} catch (e) {
const is_ok = e.message === "unreachable" && exit_code === 0;
if (!is_ok) {
console.error(e);
}
}
}
if (typeof module !== 'undefined') {
module.exports = {
roc_web_platform_run,
};
}

68
examples/hello-web/platform/host.zig Normal file
View file

@ -0,0 +1,68 @@
const std = @import("std");
const str = @import("str");
const RocStr = str.RocStr;
const testing = std.testing;
const expectEqual = testing.expectEqual;
const expect = testing.expect;
comptime {
// This is a workaround for https://github.com/ziglang/zig/issues/8218
// which is only necessary on macOS.
//
// Once that issue is fixed, we can undo the changes in
// 177cf12e0555147faa4d436e52fc15175c2c4ff0 and go back to passing
// -fcompiler-rt in link.rs instead of doing this. Note that this
// workaround is present in many host.zig files, so make sure to undo
// it everywhere!
if (std.builtin.os.tag == .macos) {
_ = @import("compiler_rt");
}
}
const Align = extern struct { a: usize, b: usize };
extern fn malloc(size: usize) callconv(.C) ?*align(@alignOf(Align)) c_void;
extern fn realloc(c_ptr: [*]align(@alignOf(Align)) u8, size: usize) callconv(.C) ?*c_void;
extern fn free(c_ptr: [*]align(@alignOf(Align)) u8) callconv(.C) void;
export fn roc_alloc(size: usize, alignment: u32) callconv(.C) ?*c_void {
_ = alignment;
return malloc(size);
}
export fn roc_realloc(c_ptr: *c_void, old_size: usize, new_size: usize, alignment: u32) callconv(.C) ?*c_void {
_ = old_size;
_ = alignment;
return realloc(@alignCast(@alignOf(Align), @ptrCast([*]u8, c_ptr)), new_size);
}
export fn roc_dealloc(c_ptr: *c_void, alignment: u32) callconv(.C) void {
_ = alignment;
free(@alignCast(@alignOf(Align), @ptrCast([*]u8, c_ptr)));
}
// NOTE roc_panic is provided in the JS file, so it can throw an exception
const mem = std.mem;
const Allocator = mem.Allocator;
extern fn roc__mainForHost_1_exposed(*RocStr) void;
const Unit = extern struct {};
extern fn js_display_roc_string(str_bytes: ?[*]u8, str_len: usize) void;
pub fn main() u8 {
// actually call roc to populate the callresult
var callresult = RocStr.empty();
roc__mainForHost_1_exposed(&callresult);
// display the result using JavaScript
js_display_roc_string(callresult.asU8ptr(), callresult.len());
callresult.deinit();
return 0;
}

25
examples/hello-web/test-node.js Normal file
View file

@ -0,0 +1,25 @@
/**
* Node.js test file for hello-web example
* We are not running this in CI currently, and Node.js is not a Roc dependency.
* But if you happen to have it, you can run this.
*/
// Node doesn't have the fetch API
const fs = require("fs/promises");
global.fetch = (filename) =>
fs.readFile(filename).then((buffer) => ({
arrayBuffer() {
return buffer;
},
}));
const { roc_web_platform_run } = require("./platform/host");
roc_web_platform_run("./hello-world.wasm", (string_from_roc) => {
const expected = "Hello, World!";
if (string_from_roc !== expected) {
console.error(`Expected "${expected}", but got "${string_from_roc}"`);
process.exit(1);
}
console.log("OK");
});

118
examples/hello-world/platform/host.c
View file

@ -1,89 +1,81 @@
#include <stdlib.h>
#include <unistd.h>
#include <stdio.h>
#include <errno.h>
#include <string.h>
#include <stdbool.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
void* roc_alloc(size_t size, unsigned int alignment) {
return malloc(size);
void* roc_alloc(size_t size, unsigned int alignment) { return malloc(size); }
void* roc_realloc(void* ptr, size_t old_size, size_t new_size,
unsigned int alignment) {
return realloc(ptr, new_size);
}
void* roc_realloc(void* ptr, size_t old_size, size_t new_size, unsigned int alignment) {
return realloc(ptr, new_size);
}
void roc_dealloc(void* ptr, unsigned int alignment) {
free(ptr);
}
void roc_dealloc(void* ptr, unsigned int alignment) { free(ptr); }
void roc_panic(void* ptr, unsigned int alignment) {
char* msg = (char *)ptr;
fprintf(stderr, "Application crashed with message\n\n %s\n\nShutting down\n", msg);
exit(0);
char* msg = (char*)ptr;
fprintf(stderr,
"Application crashed with message\n\n %s\n\nShutting down\n", msg);
exit(0);
}
void* roc_memcpy(void* dest, const void* src, size_t n) {
return memcpy(dest, src, n);
}
void* roc_memset(void* str, int c, size_t n) { return memset(str, c, n); }
struct RocStr {
char* bytes;
size_t len;
char* bytes;
size_t len;
};
bool is_small_str(struct RocStr str) {
return ((ssize_t)str.len) < 0;
}
bool is_small_str(struct RocStr str) { return ((ssize_t)str.len) < 0; }
// Determine the length of the string, taking into
// account the small string optimization
size_t roc_str_len(struct RocStr str) {
char* bytes = (char*)&str;
char last_byte = bytes[sizeof(str) - 1];
char last_byte_xored = last_byte ^ 0b10000000;
size_t small_len = (size_t)(last_byte_xored);
size_t big_len = str.len;
char* bytes = (char*)&str;
char last_byte = bytes[sizeof(str) - 1];
char last_byte_xored = last_byte ^ 0b10000000;
size_t small_len = (size_t)(last_byte_xored);
size_t big_len = str.len;
// Avoid branch misprediction costs by always
// determining both small_len and big_len,
// so this compiles to a cmov instruction.
if (is_small_str(str)) {
return small_len;
} else {
return big_len;
}
// Avoid branch misprediction costs by always
// determining both small_len and big_len,
// so this compiles to a cmov instruction.
if (is_small_str(str)) {
return small_len;
} else {
return big_len;
}
}
struct RocCallResult {
size_t flag;
struct RocStr content;
};
extern void roc__mainForHost_1_exposed(struct RocCallResult *re);
extern struct RocStr roc__mainForHost_1_exposed();
int main() {
// Make space for the Roc call result
struct RocCallResult call_result;
struct RocStr str = roc__mainForHost_1_exposed();
// Call Roc to populate call_result
roc__mainForHost_1_exposed(&call_result);
// Determine str_len and the str_bytes pointer,
// taking into account the small string optimization.
size_t str_len = roc_str_len(str);
char* str_bytes;
// Determine str_len and the str_bytes pointer,
// taking into account the small string optimization.
struct RocStr str = call_result.content;
size_t str_len = roc_str_len(str);
char* str_bytes;
if (is_small_str(str)) {
str_bytes = (char*)&str;
} else {
str_bytes = str.bytes;
}
if (is_small_str(str)) {
str_bytes = (char*)&str;
} else {
str_bytes = str.bytes;
}
// Write to stdout
if (write(1, str_bytes, str_len) >= 0) {
// Writing succeeded!
return 0;
} else {
printf("Error writing to stdout: %s\n", strerror(errno));
// Write to stdout
if (write(1, str_bytes, str_len) >= 0) {
// Writing succeeded!
return 0;
} else {
printf("Error writing to stdout: %s\n", strerror(errno));
return 1;
}
return 1;
}
}

23
examples/hello-zig/platform/host.zig
View file

@ -23,6 +23,8 @@ const Align = extern struct { a: usize, b: usize };
extern fn malloc(size: usize) callconv(.C) ?*align(@alignOf(Align)) c_void;
extern fn realloc(c_ptr: [*]align(@alignOf(Align)) u8, size: usize) callconv(.C) ?*c_void;
extern fn free(c_ptr: [*]align(@alignOf(Align)) u8) callconv(.C) void;
extern fn memcpy(dst: [*]u8, src: [*]u8, size: usize) callconv(.C) void;
extern fn memset(dst: [*]u8, value: i32, size: usize) callconv(.C) void;
export fn roc_alloc(size: usize, alignment: u32) callconv(.C) ?*c_void {
_ = alignment;
@ -51,12 +53,18 @@ export fn roc_panic(c_ptr: *c_void, tag_id: u32) callconv(.C) void {
std.process.exit(0);
}
export fn roc_memcpy(dst: [*]u8, src: [*]u8, size: usize) callconv(.C) void{
return memcpy(dst, src, size);
}
export fn roc_memset(dst: [*]u8, value: i32, size: usize) callconv(.C) void{
return memset(dst, value, size);
}
const mem = std.mem;
const Allocator = mem.Allocator;
extern fn roc__mainForHost_1_exposed(*RocCallResult) void;
const RocCallResult = extern struct { flag: u64, content: RocStr };
extern fn roc__mainForHost_1_exposed() RocStr;
const Unit = extern struct {};
@ -64,20 +72,17 @@ pub fn main() u8 {
const stdout = std.io.getStdOut().writer();
const stderr = std.io.getStdErr().writer();
// make space for the result
var callresult = RocCallResult{ .flag = 0, .content = RocStr.empty() };
// start time
var ts1: std.os.timespec = undefined;
std.os.clock_gettime(std.os.CLOCK_REALTIME, &ts1) catch unreachable;
// actually call roc to populate the callresult
roc__mainForHost_1_exposed(&callresult);
var callresult = roc__mainForHost_1_exposed();
// stdout the result
stdout.print("{s}\n", .{callresult.content.asSlice()}) catch unreachable;
stdout.print("{s}\n", .{callresult.asSlice()}) catch unreachable;
callresult.content.deinit();
callresult.deinit();
// end time
var ts2: std.os.timespec = undefined;

23
examples/quicksort/platform/host.zig
View file

@ -20,12 +20,14 @@ comptime {
const mem = std.mem;
const Allocator = mem.Allocator;
extern fn roc__mainForHost_1_exposed(RocList, *RocCallResult) void;
extern fn roc__mainForHost_1_exposed(RocList) RocList;
const Align = extern struct { a: usize, b: usize };
extern fn malloc(size: usize) callconv(.C) ?*align(@alignOf(Align)) c_void;
extern fn realloc(c_ptr: [*]align(@alignOf(Align)) u8, size: usize) callconv(.C) ?*c_void;
extern fn free(c_ptr: [*]align(@alignOf(Align)) u8) callconv(.C) void;
extern fn memcpy(dst: [*]u8, src: [*]u8, size: usize) callconv(.C) void;
extern fn memset(dst: [*]u8, value: i32, size: usize) callconv(.C) void;
const DEBUG: bool = false;
@ -67,13 +69,19 @@ export fn roc_panic(c_ptr: *c_void, tag_id: u32) callconv(.C) void {
std.process.exit(0);
}
export fn roc_memcpy(dst: [*]u8, src: [*]u8, size: usize) callconv(.C) void{
return memcpy(dst, src, size);
}
export fn roc_memset(dst: [*]u8, value: i32, size: usize) callconv(.C) void{
return memset(dst, value, size);
}
// warning! the array is currently stack-allocated so don't make this too big
const NUM_NUMS = 100;
const RocList = extern struct { elements: [*]i64, length: usize };
const RocCallResult = extern struct { flag: u64, content: RocList };
const Unit = extern struct {};
pub export fn main() u8 {
@ -93,19 +101,16 @@ pub export fn main() u8 {
const roc_list = RocList{ .elements = numbers, .length = NUM_NUMS };
// make space for the result
var callresult = RocCallResult{ .flag = 0, .content = undefined };
// start time
var ts1: std.os.timespec = undefined;
std.os.clock_gettime(std.os.CLOCK_REALTIME, &ts1) catch unreachable;
// actually call roc to populate the callresult
roc__mainForHost_1_exposed(roc_list, &callresult);
var callresult = roc__mainForHost_1_exposed(roc_list);
// stdout the result
const length = std.math.min(20, callresult.content.length);
var result = callresult.content.elements[0..length];
const length = std.math.min(20, callresult.length);
var result = callresult.elements[0..length];
for (result) |x, i| {
if (i == 0) {

33
linker/Cargo.toml Normal file
View file

@ -0,0 +1,33 @@
[package]
name = "roc_linker"
version = "0.1.0"
authors = ["The Roc Contributors"]
license = "UPL-1.0"
repository = "https://github.com/rtfeldman/roc"
edition = "2018"
description = "A surgical linker for Roc"
[lib]
name = "roc_linker"
path = "src/lib.rs"
[[bin]]
name = "link"
path = "src/main.rs"
test = false
bench = false
[dependencies]
roc_mono = { path = "../compiler/mono" }
roc_build = { path = "../compiler/build", default-features = false }
roc_collections = { path = "../compiler/collections" }
bumpalo = { version = "3.6", features = ["collections"] }
# TODO switch to clap 3.0.0 once it's out. Tried adding clap = "~3.0.0-beta.1" and cargo wouldn't accept it
clap = { git = "https://github.com/rtfeldman/clap", branch = "master" }
iced-x86 = "1.14"
memmap2 = "0.3"
object = { version = "0.26", features = ["read", "write"] }
serde = { version = "1.0", features = ["derive"] }
bincode = "1.3"
target-lexicon = "0.12.2"
tempfile = "3.1.0"

44
linker/README.md Normal file
View file

@ -0,0 +1,44 @@
# The Roc Surgical Linker
This linker has the goal of being extremely slim lined and fast.
It is focused on the scope of only linking platforms to Roc applications.
This restriction enables ignoring most of linking.
## General Overview
This linker is run in 2 phases: preprocessing and surigical linking.
### Platform Preprocessor
1. Dynamically link the platform to a dummy Roc application dynamic library
1. Create metadata related to Roc dynamically linked functions
- Symbols that need to be redefined
- Call locations that need to be modified for each symbol
- Locations of special roc functions (roc_alloc, roc_dealloc, builtins, etc)
1. Modify the main executable to no longer be dynamically link
- Delete dependency on dynamic library
- Remove symbols from the dynamic table (maybe add them to the regular table?)
- Delete GOT and PLT entries
- Remove relocations from the dynamic table
- Add extra header information about new text and data section at end of file
### Surgical Linker
1. Build off of preprocessed platform
1. Append text and data of application, dealing with app relocations
1. Surgically update all call locations in the platform
1. Surgically update call information in the application (also dealing with other relocations for builtins)
## TODO (In a lightly prioritized order)
- Run CLI tests and/or benchmarks with the Roc Linker.
- Test with an executable completely generated by Cargo (It will hopefully work out of the box like zig).
- Add Macho support
- Honestly should be almost exactly the same code.
This means we likely need to do a lot of refactoring to minimize the duplicate code.
The fun of almost but not quite the same.
- Add PE support
- As a prereq, we need roc building on Windows (I'm not sure it does currently).
- Definitely a solid bit different than elf, but hopefully after refactoring for Macho, won't be that crazy to add.
- Look at enabling completely in memory linking that could be used with `roc run` and/or `roc repl`
- Add a feature to the compiler to make this linker optional.

1626
linker/src/lib.rs Normal file
View file

File diff suppressed because it is too large Load diff

20
linker/src/main.rs Normal file
View file

@ -0,0 +1,20 @@
use roc_linker::{build_app, preprocess, surgery, CMD_PREPROCESS, CMD_SURGERY};
use std::io;
fn main() -> io::Result<()> {
let matches = build_app().get_matches();
let exit_code = match matches.subcommand_name() {
None => Ok::<i32, io::Error>(-1),
Some(CMD_PREPROCESS) => {
let sub_matches = matches.subcommand_matches(CMD_PREPROCESS).unwrap();
preprocess(sub_matches)
}
Some(CMD_SURGERY) => {
let sub_matches = matches.subcommand_matches(CMD_SURGERY).unwrap();
surgery(sub_matches)
}
_ => unreachable!(),
}?;
std::process::exit(exit_code);
}

31
linker/src/metadata.rs Normal file
View file

@ -0,0 +1,31 @@
use roc_collections::all::MutMap;
use serde::{Deserialize, Serialize};
#[derive(Serialize, Deserialize, PartialEq, Debug)]
pub struct SurgeryEntry {
pub file_offset: u64,
pub virtual_offset: u64,
pub size: u8,
}
// TODO: Reanalyze each piece of data in this struct.
// I think a number of them can be combined to reduce string duplication.
// Also I think a few of them aren't need.
// For example, I think preprocessing can deal with all shifting and remove the need for added_byte_count.
#[derive(Default, Serialize, Deserialize, PartialEq, Debug)]
pub struct Metadata {
pub app_functions: Vec<String>,
// offset followed by address.
pub plt_addresses: MutMap<String, (u64, u64)>,
pub surgeries: MutMap<String, Vec<SurgeryEntry>>,
pub dynamic_symbol_indices: MutMap<String, u64>,
pub roc_symbol_vaddresses: MutMap<String, u64>,
pub exec_len: u64,
pub load_align_constraint: u64,
pub added_byte_count: u64,
pub last_vaddr: u64,
pub dynamic_section_offset: u64,
pub dynamic_symbol_table_section_offset: u64,
pub symbol_table_section_offset: u64,
pub symbol_table_size: u64,
}

11
linker/tests/fib/.gitignore vendored Normal file
View file

@ -0,0 +1,11 @@
fib
zig-cache
zig-out
*.o
dynhost
preprocessedhost
metadata
libapp.so

Some files were not shown because too many files have changed in this diff Show more