## Summary
This PR updates the entire parser stack in multiple ways:
### Make the lexer lazy
* https://github.com/astral-sh/ruff/pull/11244
* https://github.com/astral-sh/ruff/pull/11473
Previously, Ruff's lexer would act as an iterator. The parser would
collect all the tokens in a vector first and then process the tokens to
create the syntax tree.
The first task in this project is to update the entire parsing flow to
make the lexer lazy. This includes the `Lexer`, `TokenSource`, and
`Parser`. For context, the `TokenSource` is a wrapper around the `Lexer`
to filter out the trivia tokens[^1]. Now, the parser will ask the token
source to get the next token and only then the lexer will continue and
emit the token. This means that the lexer needs to be aware of the
"current" token. When the `next_token` is called, the current token will
be updated with the newly lexed token.
The main motivation to make the lexer lazy is to allow re-lexing a token
in a different context. This is going to be really useful to make the
parser error resilience. For example, currently the emitted tokens
remains the same even if the parser can recover from an unclosed
parenthesis. This is important because the lexer emits a
`NonLogicalNewline` in parenthesized context while a normal `Newline` in
non-parenthesized context. This different kinds of newline is also used
to emit the indentation tokens which is important for the parser as it's
used to determine the start and end of a block.
Additionally, this allows us to implement the following functionalities:
1. Checkpoint - rewind infrastructure: The idea here is to create a
checkpoint and continue lexing. At a later point, this checkpoint can be
used to rewind the lexer back to the provided checkpoint.
2. Remove the `SoftKeywordTransformer` and instead use lookahead or
speculative parsing to determine whether a soft keyword is a keyword or
an identifier
3. Remove the `Tok` enum. The `Tok` enum represents the tokens emitted
by the lexer but it contains owned data which makes it expensive to
clone. The new `TokenKind` enum just represents the type of token which
is very cheap.
This brings up a question as to how will the parser get the owned value
which was stored on `Tok`. This will be solved by introducing a new
`TokenValue` enum which only contains a subset of token kinds which has
the owned value. This is stored on the lexer and is requested by the
parser when it wants to process the data. For example:
8196720f80/crates/ruff_python_parser/src/parser/expression.rs (L1260-L1262)
[^1]: Trivia tokens are `NonLogicalNewline` and `Comment`
### Remove `SoftKeywordTransformer`
* https://github.com/astral-sh/ruff/pull/11441
* https://github.com/astral-sh/ruff/pull/11459
* https://github.com/astral-sh/ruff/pull/11442
* https://github.com/astral-sh/ruff/pull/11443
* https://github.com/astral-sh/ruff/pull/11474
For context,
https://github.com/RustPython/RustPython/pull/4519/files#diff-5de40045e78e794aa5ab0b8aacf531aa477daf826d31ca129467703855408220
added support for soft keywords in the parser which uses infinite
lookahead to classify a soft keyword as a keyword or an identifier. This
is a brilliant idea as it basically wraps the existing Lexer and works
on top of it which means that the logic for lexing and re-lexing a soft
keyword remains separate. The change here is to remove
`SoftKeywordTransformer` and let the parser determine this based on
context, lookahead and speculative parsing.
* **Context:** The transformer needs to know the position of the lexer
between it being at a statement position or a simple statement position.
This is because a `match` token starts a compound statement while a
`type` token starts a simple statement. **The parser already knows
this.**
* **Lookahead:** Now that the parser knows the context it can perform
lookahead of up to two tokens to classify the soft keyword. The logic
for this is mentioned in the PR implementing it for `type` and `match
soft keyword.
* **Speculative parsing:** This is where the checkpoint - rewind
infrastructure helps. For `match` soft keyword, there are certain cases
for which we can't classify based on lookahead. The idea here is to
create a checkpoint and keep parsing. Based on whether the parsing was
successful and what tokens are ahead we can classify the remaining
cases. Refer to #11443 for more details.
If the soft keyword is being parsed in an identifier context, it'll be
converted to an identifier and the emitted token will be updated as
well. Refer
8196720f80/crates/ruff_python_parser/src/parser/expression.rs (L487-L491).
The `case` soft keyword doesn't require any special handling because
it'll be a keyword only in the context of a match statement.
### Update the parser API
* https://github.com/astral-sh/ruff/pull/11494
* https://github.com/astral-sh/ruff/pull/11505
Now that the lexer is in sync with the parser, and the parser helps to
determine whether a soft keyword is a keyword or an identifier, the
lexer cannot be used on its own. The reason being that it's not
sensitive to the context (which is correct). This means that the parser
API needs to be updated to not allow any access to the lexer.
Previously, there were multiple ways to parse the source code:
1. Passing the source code itself
2. Or, passing the tokens
Now that the lexer and parser are working together, the API
corresponding to (2) cannot exists. The final API is mentioned in this
PR description: https://github.com/astral-sh/ruff/pull/11494.
### Refactor the downstream tools (linter and formatter)
* https://github.com/astral-sh/ruff/pull/11511
* https://github.com/astral-sh/ruff/pull/11515
* https://github.com/astral-sh/ruff/pull/11529
* https://github.com/astral-sh/ruff/pull/11562
* https://github.com/astral-sh/ruff/pull/11592
And, the final set of changes involves updating all references of the
lexer and `Tok` enum. This was done in two-parts:
1. Update all the references in a way that doesn't require any changes
from this PR i.e., it can be done independently
* https://github.com/astral-sh/ruff/pull/11402
* https://github.com/astral-sh/ruff/pull/11406
* https://github.com/astral-sh/ruff/pull/11418
* https://github.com/astral-sh/ruff/pull/11419
* https://github.com/astral-sh/ruff/pull/11420
* https://github.com/astral-sh/ruff/pull/11424
2. Update all the remaining references to use the changes made in this
PR
For (2), there were various strategies used:
1. Introduce a new `Tokens` struct which wraps the token vector and add
methods to query a certain subset of tokens. These includes:
1. `up_to_first_unknown` which replaces the `tokenize` function
2. `in_range` and `after` which replaces the `lex_starts_at` function
where the former returns the tokens within the given range while the
latter returns all the tokens after the given offset
2. Introduce a new `TokenFlags` which is a set of flags to query certain
information from a token. Currently, this information is only limited to
any string type token but can be expanded to include other information
in the future as needed. https://github.com/astral-sh/ruff/pull/11578
3. Move the `CommentRanges` to the parsed output because this
information is common to both the linter and the formatter. This removes
the need for `tokens_and_ranges` function.
## Test Plan
- [x] Update and verify the test snapshots
- [x] Make sure the entire test suite is passing
- [x] Make sure there are no changes in the ecosystem checks
- [x] Run the fuzzer on the parser
- [x] Run this change on dozens of open-source projects
### Running this change on dozens of open-source projects
Refer to the PR description to get the list of open source projects used
for testing.
Now, the following tests were done between `main` and this branch:
1. Compare the output of `--select=E999` (syntax errors)
2. Compare the output of default rule selection
3. Compare the output of `--select=ALL`
**Conclusion: all output were same**
## What's next?
The next step is to introduce re-lexing logic and update the parser to
feed the recovery information to the lexer so that it can emit the
correct token. This moves us one step closer to having error resilience
in the parser and provides Ruff the possibility to lint even if the
source code contains syntax errors.
## Summary
Follow up to https://github.com/astral-sh/ruff/pull/11521
Removes the extra added complexity for catch all match cases. This
matches the implementation of plain `else` statements.
## Test Plan
Added new test cases.
---------
Co-authored-by: Dhruv Manilawala <dhruvmanila@gmail.com>
## Summary
This moves the string-prefix enumerations in `ruff_python_ast` to a
separate submodule. I think this helps clarify that these prefixes are
purely abstract: they only depend on each other, and do not depend on
any of the other code in `nodes.rs` in any way. Moreover, while various
AST nodes _use_ them, they're not really nodes themselves, so they feel
slightly out of place in `nodes.rs`.
I considered moving all of them to `str.rs`, but it felt like enough
code that it could be a separate submodule.
## Test Plan
`cargo test`
## Summary
This PR adds a newtype wrapper around `Vec<FStringElement>` that derefs
to a `&Vec<FStringElement>`.
Both f-string and format specifier are made up of `Vec<FStringElement>`.
By creating a newtype wrapper around it, we can share the methods for
both parent types.
## Summary
This PR adds support to iterate over each part of a string-like
expression.
This similar to the one in the formatter:
128414cd95/crates/ruff_python_formatter/src/string/any.rs (L121-L125)
Although I don't think it's a 1-1 replacement in the formatter because
the one implemented in the formatter has another information for certain
variants (as can be seen for `FString`).
The main motivation for this is to avoid duplication for rules which
work only on the parts of the string and doesn't require any information
from the parent node. Here, the parent node being the expression node
which could be an implicitly concatenated string.
This PR also updates certain rule implementation to make use of this and
avoids logic duplication.
## Summary
This PR renames `AnyStringKind` to `AnyStringFlags` and `AnyStringFlags`
to `AnyStringFlagsInner`.
The main motivation is to have consistent usage of "kind" and "flags".
For each string kind, it's "flags" like `StringLiteralFlags`,
`BytesLiteralFlags`, and `FStringFlags` but it was `AnyStringKind` for
the "any" variant.
## Summary
This PR removes the cyclic dev dependency some of the crates had with
the parser crate.
The cyclic dependencies are:
* `ruff_python_ast` has a **dev dependency** on `ruff_python_parser` and
`ruff_python_parser` directly depends on `ruff_python_ast`
* `ruff_python_trivia` has a **dev dependency** on `ruff_python_parser`
and `ruff_python_parser` has an indirect dependency on
`ruff_python_trivia` (`ruff_python_parser` - `ruff_python_ast` -
`ruff_python_trivia`)
Specifically, this PR does the following:
* Introduce two new crates
* `ruff_python_ast_integration_tests` and move the tests from the
`ruff_python_ast` crate which uses the parser in this crate
* `ruff_python_trivia_integration_tests` and move the tests from the
`ruff_python_trivia` crate which uses the parser in this crate
### Motivation
The main motivation for this PR is to help development. Before this PR,
`rust-analyzer` wouldn't provide any intellisense in the
`ruff_python_parser` crate regarding the symbols in `ruff_python_ast`
crate.
```
[ERROR][2024-05-03 13:47:06] .../vim/lsp/rpc.lua:770 "rpc" "/Users/dhruv/.cargo/bin/rust-analyzer" "stderr" "[ERROR project_model::workspace] cyclic deps: ruff_python_parser(Idx::<CrateData>(50)) -> ruff_python_ast(Idx::<CrateData>(37)), alternative path: ruff_python_ast(Idx::<CrateData>(37)) -> ruff_python_parser(Idx::<CrateData>(50))\n"
```
## Test Plan
Check the logs of `rust-analyzer` to not see any signs of cyclic
dependency.
## Summary
This PR removes the `ImportMap` implementation and all its routing
through ruff.
The import map was added in https://github.com/astral-sh/ruff/pull/3243
but we then never ended up using it to do cross file analysis.
We are now working on adding multifile analysis to ruff, and revisit
import resolution as part of it.
```
hyperfine --warmup 10 --runs 20 --setup "./target/release/ruff clean" \
"./target/release/ruff check crates/ruff_linter/resources/test/cpython -e -s --extend-select=I" \
"./target/release/ruff-import check crates/ruff_linter/resources/test/cpython -e -s --extend-select=I"
Benchmark 1: ./target/release/ruff check crates/ruff_linter/resources/test/cpython -e -s --extend-select=I
Time (mean ± σ): 37.6 ms ± 0.9 ms [User: 52.2 ms, System: 63.7 ms]
Range (min … max): 35.8 ms … 39.8 ms 20 runs
Benchmark 2: ./target/release/ruff-import check crates/ruff_linter/resources/test/cpython -e -s --extend-select=I
Time (mean ± σ): 36.0 ms ± 0.7 ms [User: 50.3 ms, System: 58.4 ms]
Range (min … max): 34.5 ms … 37.6 ms 20 runs
Summary
./target/release/ruff-import check crates/ruff_linter/resources/test/cpython -e -s --extend-select=I ran
1.04 ± 0.03 times faster than ./target/release/ruff check crates/ruff_linter/resources/test/cpython -e -s --extend-select=I
```
I suspect that the performance improvement should even be more
significant for users that otherwise don't have any diagnostics.
```
hyperfine --warmup 10 --runs 20 --setup "cd ../ecosystem/airflow && ../../ruff/target/release/ruff clean" \
"./target/release/ruff check ../ecosystem/airflow -e -s --extend-select=I" \
"./target/release/ruff-import check ../ecosystem/airflow -e -s --extend-select=I"
Benchmark 1: ./target/release/ruff check ../ecosystem/airflow -e -s --extend-select=I
Time (mean ± σ): 53.7 ms ± 1.8 ms [User: 68.4 ms, System: 63.0 ms]
Range (min … max): 51.1 ms … 58.7 ms 20 runs
Benchmark 2: ./target/release/ruff-import check ../ecosystem/airflow -e -s --extend-select=I
Time (mean ± σ): 50.8 ms ± 1.4 ms [User: 50.7 ms, System: 60.9 ms]
Range (min … max): 48.5 ms … 55.3 ms 20 runs
Summary
./target/release/ruff-import check ../ecosystem/airflow -e -s --extend-select=I ran
1.06 ± 0.05 times faster than ./target/release/ruff check ../ecosystem/airflow -e -s --extend-select=I
```
## Test Plan
`cargo test`
## Summary
This PR adds a basic README for the `ruff_python_parser` crate and
updates the CONTRIBUTING docs with the fuzzer and benchmark section.
Additionally, it also updates some inline documentation within the
parser crate and splits the `parse_program` function into
`parse_single_expression` and `parse_module` which will be called by
matching against the `Mode`.
This PR doesn't go into too much internal detail around the parser logic
due to the following reasons:
1. Where should the docs go? Should it be as a module docs in `lib.rs`
or in README?
2. The parser is still evolving and could include a lot of refactors
with the future work (feedback loop and improved error recovery and
resilience)
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
## Summary
I happened to notice that we box `TypeParams` on `StmtClassDef` but not
on `StmtFunctionDef` and wondered why, since `StmtFunctionDef` is bigger
and sets the size of `Stmt`.
@charliermarsh found that at the time we started boxing type params on
classes, classes were the largest statement type (see #6275), but that's
no longer true.
So boxing type-params also on functions reduces the overall size of
`Stmt`.
## Test Plan
The `<=` size tests are a bit irritating (since their failure doesn't
tell you the actual size), but I manually confirmed that the size is
actually 120 now.
(Supersedes #9152, authored by @LaBatata101)
## Summary
This PR replaces the current parser generated from LALRPOP to a
hand-written recursive descent parser.
It also updates the grammar for [PEP
646](https://peps.python.org/pep-0646/) so that the parser outputs the
correct AST. For example, in `data[*x]`, the index expression is now a
tuple with a single starred expression instead of just a starred
expression.
Beyond the performance improvements, the parser is also error resilient
and can provide better error messages. The behavior as seen by any
downstream tools isn't changed. That is, the linter and formatter can
still assume that the parser will _stop_ at the first syntax error. This
will be updated in the following months.
For more details about the change here, refer to the PR corresponding to
the individual commits and the release blog post.
## Test Plan
Write _lots_ and _lots_ of tests for both valid and invalid syntax and
verify the output.
## Acknowledgements
- @MichaReiser for reviewing 100+ parser PRs and continuously providing
guidance throughout the project
- @LaBatata101 for initiating the transition to a hand-written parser in
#9152
- @addisoncrump for implementing the fuzzer which helped
[catch](https://github.com/astral-sh/ruff/pull/10903)
[a](https://github.com/astral-sh/ruff/pull/10910)
[lot](https://github.com/astral-sh/ruff/pull/10966)
[of](https://github.com/astral-sh/ruff/pull/10896)
[bugs](https://github.com/astral-sh/ruff/pull/10877)
---------
Co-authored-by: Victor Hugo Gomes <labatata101@linuxmail.org>
Co-authored-by: Micha Reiser <micha@reiser.io>
## Summary
Historically, given:
```python
__all__ = [ # noqa: F822
"Bernoulli",
"Beta",
"Binomial",
]
```
The F822 violations would be attached to the `__all__`, so this `# noqa`
would be enforced for _all_ definitions in the list. This changed in
https://github.com/astral-sh/ruff/pull/10525 for the better, in that we
now use the range of each string. But these `# noqa` directives stopped
working.
This PR sets the `__all__` as a parent range in the diagnostic, so that
these directives are respected once again.
Closes https://github.com/astral-sh/ruff/issues/10795.
## Test Plan
`cargo test`
## Summary
This PR adds the `as_str` implementation for all the operator methods.
It already exists for `CmpOp` which is being [used in the
linter](ffcd77860c/crates/ruff_linter/src/rules/flake8_simplify/rules/key_in_dict.rs (L117))
and it makes sense to implement it for the rest as well. This will also
be utilized in error messages for the new parser.
## Summary
This PR removes unused operator methods and impl traits. There is
already the `is_macro::Is` implementation for all the operators and this
seems unnecessary.
## Summary
This PR updates the `StringLike::FString` variant to use `ExprFString`
instead of `FStringLiteralElement`.
For context, the reason it used `FStringLiteralElement` is that the node
is actually the string part of an f-string ("foo" in `f"foo{x}"`). But,
this is inconsistent with other variants where the captured value is the
_entire_ string.
This is also problematic w.r.t. implicitly concatenated strings. Any
rules which work with `StringLike::FString` doesn't account for the
string part in an implicitly concatenated f-strings. For example, we
don't flag confusable character in the first part of `"𝐁ad" f"𝐁ad
string"`, but only the second part
(https://play.ruff.rs/16071c4c-a1dd-4920-b56f-e2ce2f69c843).
### Update `PYI053`
_This is included in this PR because otherwise it requires a temporary
workaround to be compatible with the old logic._
This PR also updates the `PYI053` (`string-or-bytes-too-long`) rule for
f-string to consider _all_ the visible characters in a f-string,
including the ones which are implicitly concatenated. This is consistent
with implicitly concatenated strings and bytes.
For example,
```python
def foo(
# We count all the characters here
arg1: str = '51 character ' 'stringgggggggggggggggggggggggggggggggg',
# But not here because of the `{x}` replacement field which _breaks_ them up into two chunks
arg2: str = f'51 character {x} stringgggggggggggggggggggggggggggggggggggggggggggg',
) -> None: ...
```
This PR fixes it to consider all _visible_ characters inside an f-string
which includes expressions as well.
fixes: #10310fixes: #10307
## Test Plan
Add new test cases and update the snapshots.
## Review
To facilitate the review process, the change have been split into two
commits: one which has the code change while the other has the test
cases and updated snapshots.
## Summary
This PR adds methods on `FString` to iterate over the two different kind
of elements it can have - literals and expressions. This is similar to
the methods we have on `ExprFString`.
---------
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
This PR modifies our AST so that nodes for string literals, bytes literals and f-strings all retain the following information:
- The quoting style used (double or single quotes)
- Whether the string is triple-quoted or not
- Whether the string is raw or not
This PR is a followup to #10256. Like with that PR, this PR does not, in itself, fix any bugs. However, it means that we will have the necessary information to preserve quoting style and rawness of strings in the `ExprGenerator` in a followup PR, which will allow us to provide a fix for https://github.com/astral-sh/ruff/issues/7799.
The information is recorded on the AST nodes using a bitflag field on each node, similarly to how we recorded the information on `Tok::String`, `Tok::FStringStart` and `Tok::FStringMiddle` tokens in #10298. Rather than reusing the bitflag I used for the tokens, however, I decided to create a custom bitflag for each AST node.
Using different bitflags for each node allows us to make invalid states unrepresentable: it is valid to set a `u` prefix on a string literal, but not on a bytes literal or an f-string. It also allows us to have better debug representations for each AST node modified in this PR.
## Summary
When you try to remove an internal representation leaking into another
type and end up rewriting a simple version of `smallvec`.
The goal of this PR is to replace the `Box<[&'a str]>` with
`Box<QualifiedName>` to avoid that the internal `QualifiedName`
representation leaks (and it gives us a nicer API too). However, doing
this when `QualifiedName` uses `SmallVec` internally gives us all sort
of funny lifetime errors. I was lost but @BurntSushi came to rescue me.
He figured out that `smallvec` has a variance problem which is already
tracked in https://github.com/servo/rust-smallvec/issues/146
To fix the variants problem, I could use the smallvec-2-alpha-4 or
implement our own smallvec. I went with implementing our own small vec
for this specific problem. It obviously isn't as sophisticated as
smallvec (only uses safe code), e.g. it doesn't perform any size
optimizations, but it does its job.
Other changes:
* Removed `Imported::qualified_name` (the version that returns a
`String`). This can be replaced by calling `ToString` on the qualified
name.
* Renamed `Imported::call_path` to `qualified_name` and changed its
return type to `&QualifiedName`.
* Renamed `QualifiedName::imported` to `user_defined` which is the more
common term when talking about builtins vs the rest/user defined
functions.
## Test plan
`cargo test`
The expression types in our AST are called `ExprYield`, `ExprAwait`,
`ExprStringLiteral` etc, except `ExprNamedExpr`, `ExprIfExpr` and
`ExprGenratorExpr`. This seems to align with [Python AST's
naming](https://docs.python.org/3/library/ast.html) but feels
inconsistent and excessive.
This PR removes the `Expr` postfix from `ExprNamedExpr`, `ExprIfExpr`,
and `ExprGeneratorExpr`.
