Commit graph

2255 commits

Author SHA1 Message Date
Charlie Marsh
10cd6b94c9
Normalize extras in lockfile (#3958)
## Summary

Previously, when we locked something like `flask[dotenv]`, we created
two separate distributions in the lockfile: one for `flask`, which
included the base dependencies, and one for `flask[dotenv]`, which
included the base dependencies _and_ the `dotenv` dependencies. This was
easy to implement, but it meant that we were duplicating all of the
distribution files for every extra, and duplicating all of the base
dependencies for every extra.

This PR normalizes the data such that we now have one entry per
distribution (i.e., `ExtraName` was removed from `DistributionId`), with
an optional dependencies table with an entry per extra, like:

```toml
[[distribution]]
name = "project"
version = "0.1.0"
source = "editable+file://[TEMP_DIR]/"
sdist = { url = "file://[TEMP_DIR]/" }

[[distribution.dependencies]]
name = "anyio"
version = "3.7.0"
source = "registry+https://pypi.org/simple"

[distribution.optional-dependencies]

[[distribution.optional-dependencies.test]]
name = "iniconfig"
version = "2.0.0"
source = "registry+https://pypi.org/simple"
```

This requires a bit more work upfront, because we now need to merge
multiple packages from the `PetGraph` representation when creating the
lockfile.

Closes https://github.com/astral-sh/uv/issues/3916.
2024-06-03 19:00:35 +00:00
Charlie Marsh
362b00cc12
Remove need to return Python version in get_dependencies (#3993)
## Summary

Once we use a _range_ rather than a precise version, it won't actually
make sense to return a version here. It's no longer required, so I'm
removing it.
2024-06-03 18:42:38 +00:00
Charlie Marsh
1a60368ce4
Use PubGrubPython type in Python incompatibility reporting (#3992)
## Summary

Rather than re-testing compatibility, I think we can just rely on the
types directly.
2024-06-03 14:32:22 -04:00
Charlie Marsh
a589ad5066
Bump version to v0.2.6 (#3991) 2024-06-03 17:40:28 +00:00
Ibraheem Ahmed
1ffe18d861
Avoid race condition in OnceMap (#3987)
## Summary

Fixes a race condition in `OnceMap::wait_blocking` where the inserted
value could potentially be missed, leading to a deadlock. Fairly certain
this will resolve https://github.com/astral-sh/uv/issues/3724.
2024-06-03 12:25:58 -04:00
Charlie Marsh
29ea5d5d9a
Fix reference to --python-version patch behavior (#3989)
## Summary

We changed this at some point but didn't update the documentation.
2024-06-03 15:34:12 +00:00
konsti
8c3a8bf026
Update pubgrub to d4795a31be17669aba11eb741b4a9086acc3eb11 (#3988)
I trimmed down the diff of our pubgrub fork with upstream and
[d4795a31be17669aba11eb741b4a9086acc3eb11](d4795a31be)
(diff:
https://github.com/pubgrub-rs/pubgrub/compare/dev...astral-sh:pubgrub:perma-35)
is the result.
2024-06-03 15:33:34 +00:00
Charlie Marsh
650638fa52
Remove unstable uv lock from pip interface (#3970)
## Summary

I think we can start using `uv lock` and `uv sync` to test this instead.
2024-06-03 15:14:11 +00:00
Andrew Gallant
fbf562d5a8
uv {lock,sync}: propagate index URLs to registry client (#3986)
Otherwise the `uv lock` command wasn't respecting the index URL option.

This is a follow-up to #3984, and I believe should now allow #3970 to be
merged.
2024-06-03 14:38:36 +00:00
Charlie Marsh
a4c73fd07e
Extract index arguments into their own Clap group (#3985)
## Summary

Just removes a lot of repeated documentation. We can expand this
strategy in the future.
2024-06-03 14:27:35 +00:00
Charlie Marsh
da1782808a
Add index URL parameters to Project CLI (#3984)
## Summary

For now, this exists to facilitate testing. We may remove it in the
future in favor of other APIs.
2024-06-03 10:17:41 -04:00
Tim de Jager
1b1600c40e
feat: add back the use of extra env vars to the build dispatch (#3981)
Seems like a recent Pull removed this, couldn't directly find out which.
I'm adding it back as we rely on this API, and I do not see another way
of accessing this, or am I mistaken?

Thanks!
2024-06-03 09:13:44 -04:00
Di-Is
5c776939d2
Add override namespace to pyproject.toml/uv.toml (#3839)
<!--
Thank you for contributing to uv! To help us out with reviewing, please
consider the following:

- Does this pull request include a summary of the change? (See below.)
- Does this pull request include a descriptive title?
- Does this pull request include references to any relevant issues?
-->

## Summary

See #3834 .

This PR adds a new namespace, `override-dependencies`, to
pyproject.toml/uv.toml.
This namespace assumes that the dependencies you want to override are
written in the form of `requirements.txt`.


a example of pyproject.toml
```toml
[project]
name = "example"
version = "0.0.0"
dependencies = [
  "flask==3.0.0"
]

[tool.uv]
override-dependencies = [
  "werkzeug==2.3.0"
]
```

This will improve usability by allowing you to override dependencies
without having to specify the --override option when running `uv pip
compile/install`.

## Test Plan

added test to `crates/uv/tests/pip_compile.rs`.

---------

Co-authored-by: konstin <konstin@mailbox.org>
2024-06-03 12:15:51 +02:00
T-256
1b769b054c
Windows: Statically linked C runtime (#3966) 2024-06-03 09:18:07 +02:00
renovate[bot]
314cdbb15b
Update Rust crate axoupdater to v0.6.5 (#3972) 2024-06-02 21:52:35 -04:00
renovate[bot]
95d9bb5ee7
Update Rust crate tokio to v1.38.0 (#3976) 2024-06-03 01:51:39 +00:00
renovate[bot]
220d71ba7e
Update pre-commit hook astral-sh/ruff-pre-commit to v0.4.7 (#3974) 2024-06-03 01:46:28 +00:00
renovate[bot]
e4a9366586
Update Rust crate async-compression to v0.4.11 (#3971) 2024-06-02 21:45:41 -04:00
renovate[bot]
0addc10fca
Update Rust crate hyper-util to v0.1.5 (#3973) 2024-06-03 01:45:29 +00:00
Charlie Marsh
a23ca5bab7
Use universal resolution in uv lock (#3969)
## Summary

Wires up the optional markers in resolution, and adds
respecting-the-markers to `Lock:: to_resolution`.
2024-06-02 21:33:18 -04:00
Charlie Marsh
c500b78936
Avoid re-adding solutions to forked state (#3967)
## Summary

Running a resolution that required forking was failing due to breaking
an invariant in PubGrub. It looks like we were adding the same
incompatibility multiple times, or something like that. The issue
appears to be that when forking, we modify the current state, then clone
it as the "next state", then push to the "forked states" -- but that
means we're cloning the _modified_ state.

This PR changes the order of operations such that we clone, then modify.
It shouldn't introduce any additional clones though.
2024-06-02 17:58:25 -04:00
konsti
01d1a39c21
Add uv run --package (#3864)
Add a `--package` option that allows switching the current project in
the workspace. Wherever you are in a workspace, you should be able to
run with any other project as root. This is the uv equivalent of `cargo
run -p`.

I don't love the `--package` name, esp. since `-p` is already taken and
in general to many things start with p already.

Part of this change is moving the workspace discovery of
`ProjectWorkspace` to `Workspace` itself.

## Usage

In albatross-virtual-workspace:

```console
$ uv venv
$ uv run --preview --package bird-feeder python -c "import albatross"
   Built file:///home/konsti/projects/uv/scripts/workspaces/albatross-virtual-workspace/packages/bird-feeder
   Built file:///home/konsti/projects/uv/scripts/workspaces/albatross-virtual-workspace/packages/seeds
Built 2 editables in 167ms
Resolved 5 packages in 4ms
Installed 5 packages in 1ms
 + anyio==4.4.0
 + bird-feeder==1.0.0 (from file:///home/konsti/projects/uv/scripts/workspaces/albatross-virtual-workspace/packages/bird-feeder)
 + idna==3.6
 + seeds==1.0.0 (from file:///home/konsti/projects/uv/scripts/workspaces/albatross-virtual-workspace/packages/seeds)
 + sniffio==1.3.1
Traceback (most recent call last):
  File "<string>", line 1, in <module>
ModuleNotFoundError: No module named 'albatross'
$ uv venv
$ uv run --preview --package albatross python -c "import albatross"
   Built file:///home/konsti/projects/uv/scripts/workspaces/albatross-virtual-workspace/packages/albatross
   Built file:///home/konsti/projects/uv/scripts/workspaces/albatross-virtual-workspace/packages/bird-feeder
   Built file:///home/konsti/projects/uv/scripts/workspaces/albatross-virtual-workspace/packages/seeds
Built 3 editables in 173ms
Resolved 7 packages in 6ms
Installed 7 packages in 1ms
 + albatross==0.1.0 (from file:///home/konsti/projects/uv/scripts/workspaces/albatross-virtual-workspace/packages/albatross)
 + anyio==4.4.0
 + bird-feeder==1.0.0 (from file:///home/konsti/projects/uv/scripts/workspaces/albatross-virtual-workspace/packages/bird-feeder)
 + idna==3.6
 + seeds==1.0.0 (from file:///home/konsti/projects/uv/scripts/workspaces/albatross-virtual-workspace/packages/seeds)
 + sniffio==1.3.1
 + tqdm==4.66.4
```

In albatross-root-workspace:

```console
$ uv venv
$ uv run --preview --package bird-feeder python -c "import albatross"
  Using Python 3.12.3 interpreter at: /home/konsti/.local/bin/python3
  Creating virtualenv at: .venv
  Activate with: source .venv/bin/activate
      Finished `dev` profile [unoptimized + debuginfo] target(s) in 0.10s
       Running `/home/konsti/projects/uv/target/debug/uv run --preview --package bird-feeder python -c 'import albatross'`
     Built file:///home/konsti/projects/uv/scripts/workspaces/albatross-root-workspace/packages/bird-feeder
     Built file:///home/konsti/projects/uv/scripts/workspaces/albatross-root-workspace/packages/seeds                                              Built 2 editables in 161ms
  Resolved 5 packages in 4ms
  Installed 5 packages in 1ms
   + anyio==4.4.0
   + bird-feeder==1.0.0 (from file:///home/konsti/projects/uv/scripts/workspaces/albatross-root-workspace/packages/bird-feeder)
   + idna==3.6
   + seeds==1.0.0 (from file:///home/konsti/projects/uv/scripts/workspaces/albatross-root-workspace/packages/seeds)
   + sniffio==1.3.1
  Traceback (most recent call last):
    File "<string>", line 1, in <module>
  ModuleNotFoundError: No module named 'albatross'
$ uv venv
$ cargo run run --preview --package albatross python -c "import albatross"
Using Python 3.12.3 interpreter at: /home/konsti/.local/bin/python3
Creating virtualenv at: .venv
Activate with: source .venv/bin/activate
    Finished `dev` profile [unoptimized + debuginfo] target(s) in 0.13s
     Running `/home/konsti/projects/uv/target/debug/uv run --preview --package albatross python -c 'import albatross'`
   Built file:///home/konsti/projects/uv/scripts/workspaces/albatross-root-workspace
   Built file:///home/konsti/projects/uv/scripts/workspaces/albatross-root-workspace/packages/bird-feeder
   Built file:///home/konsti/projects/uv/scripts/workspaces/albatross-root-workspace/packages/seeds
Built 3 editables in 168ms
Resolved 7 packages in 5ms
Installed 7 packages in 1ms
 + albatross==0.1.0 (from file:///home/konsti/projects/uv/scripts/workspaces/albatross-root-workspace)
 + anyio==4.4.0
 + bird-feeder==1.0.0 (from file:///home/konsti/projects/uv/scripts/workspaces/albatross-root-workspace/packages/bird-feeder)
 + idna==3.6
 + seeds==1.0.0 (from file:///home/konsti/projects/uv/scripts/workspaces/albatross-root-workspace/packages/seeds)
 + sniffio==1.3.1
 + tqdm==4.66.4
```
2024-06-02 21:42:14 +00:00
Matthieu Darbois
0d0308c531
fix: add missing ppc64le alias for powerpc64le (#3963)
## Summary

Same as #3899 but for ppc64le, there were no tests added there so I
wouldn't know where to begin to do so.

Using image docker image `quay.io/pypa/manylinux2014_ppc64le` (uv 0.2.4)

```
[root@e3ff544d1337 ~]# python3.12 -V
Python 3.12.3
[root@e3ff544d1337 ~]# RUST_LOG="trace" uv venv py312 --python python3.12
DEBUG Searching for Python 3.12 in search path
TRACE Searching PATH for executables: python3.12, python3, python
TRACE Checking `PATH` directory for interpreters: /opt/rh/devtoolset-10/root/usr/bin
TRACE Checking `PATH` directory for interpreters: /usr/local/sbin
TRACE Checking `PATH` directory for interpreters: /usr/local/bin
TRACE Found possible Python executable: /usr/local/bin/python3.12
TRACE Querying interpreter executable at /usr/local/bin/python3.12
TRACE Querying Python at `/usr/local/bin/python3.12` did not return the expected data
unknown variant `ppc64le`, expected one of `aarch64`, `armv6l`, `armv7l`, `powerpc64le`, `powerpc64`, `x86`, `x86_64`, `s390x`
--- stdout:
{"result": "success", "markers": {"implementation_name": "cpython", "implementation_version": "3.12.3", "os_name": "posix", "platform_machine": "ppc64le", "platform_python_implementation": "CPython", "platform_release": "6.6.26-linuxkit", "platform_system": "Linux", "platform_version": "#1 SMP Sat Apr 27 04:13:19 UTC 2024", "python_full_version": "3.12.3", "python_version": "3.12", "sys_platform": "linux"}, "base_prefix": "/opt/_internal/cpython-3.12.3", "base_exec_prefix": "/opt/_internal/cpython-3.12.3", "prefix": "/opt/_internal/cpython-3.12.3", "base_executable": "/usr/local/bin/python3.12", "sys_executable": "/usr/local/bin/python3.12", "sys_path": ["/root/.cache/uv/.tmpBnM4PN", "/opt/_internal/cpython-3.12.3/lib/python312.zip", "/opt/_internal/cpython-3.12.3/lib/python3.12", "/opt/_internal/cpython-3.12.3/lib/python3.12/lib-dynload", "/opt/_internal/cpython-3.12.3/lib/python3.12/site-packages"], "stdlib": "/opt/_internal/cpython-3.12.3/lib/python3.12", "scheme": {"platlib": "/opt/_internal/cpython-3.12.3/lib/python3.12/site-packages", "purelib": "/opt/_internal/cpython-3.12.3/lib/python3.12/site-packages", "include": "/opt/_internal/cpython-3.12.3/include/python3.12", "scripts": "/opt/_internal/cpython-3.12.3/bin", "data": "/opt/_internal/cpython-3.12.3"}, "virtualenv": {"purelib": "lib/python3.12/site-packages", "platlib": "lib/python3.12/site-packages", "include": "include/site/python3.12", "scripts": "bin", "data": ""}, "platform": {"os": {"name": "manylinux", "major": 2, "minor": 17}, "arch": "ppc64le"}, "gil_disabled": false, "pointer_size": "64"}
--- stderr:

---
TRACE Skipping bad interpreter at /usr/local/bin/python3.12
TRACE Checking `PATH` directory for interpreters: /usr/sbin
TRACE Checking `PATH` directory for interpreters: /usr/bin
TRACE Found possible Python executable: /usr/bin/python
TRACE Querying interpreter executable at /usr/bin/python
TRACE Can't use Python at `/usr/bin/python`
TRACE Skipping bad interpreter at /usr/bin/python
TRACE Checking `PATH` directory for interpreters: /sbin
TRACE Checking `PATH` directory for interpreters: /bin
TRACE Found possible Python executable: /bin/python
TRACE Querying interpreter executable at /bin/python
TRACE Can't use Python at `/bin/python`
TRACE Skipping bad interpreter at /bin/python
  × No interpreter found for Python 3.12 in search path
```

---------

Co-authored-by: Charlie Marsh <crmarsh416@gmail.com>
2024-06-02 13:15:24 -04:00
konsti
1eb968fa73
Draft for user docs for workspaces (#3866)
These docs are not yet meant to be consumed by end users, for now they
allow sharing the state of workspace support in terms of features that
are relevant to users, and to gather feedback on the design. We can
progressively update this doc as workspace support matures.
2024-06-01 16:13:11 -04:00
Charlie Marsh
11324646cb
Remove some anyhow usages (#3962) 2024-06-01 20:11:23 +00:00
Charlie Marsh
a70e33d947
Move reference check into uv-git (#3961) 2024-06-01 16:02:25 -04:00
Charlie Marsh
c04a95e037
Respect resolved Git SHAs in uv lock (#3956)
## Summary

This PR ensures that if a lockfile already contains a resolved reference
(e.g., you locked with `main` previously, and it locked to a specific
commit), and you run `uv lock`, we use the same SHA, even if it's not
the latest SHA for that tag. This avoids upgrading Git dependencies
without `--upgrade`.

Closes #3920.
2024-06-01 12:40:11 +00:00
Charlie Marsh
b7d77c04cc
Add Git resolver in lieu of static hash map (#3954)
## Summary

This PR removes the static resolver map:

```rust
static RESOLVED_GIT_REFS: Lazy<Mutex<FxHashMap<RepositoryReference, GitSha>>> =
    Lazy::new(Mutex::default);
```

With a `GitResolver` struct that we now pass around on the
`BuildContext`. There should be no behavior changes here; it's purely an
internal refactor with an eye towards making it cleaner for us to
"pre-populate" the list of resolved SHAs.
2024-05-31 22:44:42 -04:00
Charlie Marsh
a0652921fc
Don't enforce that requirements files are "files" (#3947)
## Summary

This was just an opportunistic guard but it broke some use-cases.

Closes #3944.
2024-05-31 21:32:36 +00:00
Charlie Marsh
7b7da803bb
Unify editable and unnamed URL parsing (#3946)
## Summary

This will help prevent bugs like #3934 by unifying the implementations
for editables and non-editable unnamed requirements. Specifically, both
of these now go through the same parsing paths and use the same struct
representations (with the exception that the editable flag is flipped in
the first case):

```
-e ./foo/bar
./foo/bar
```

We also now support PEP 508 in editable URLs. It turns out this is just
a limitation in pip, so it's correct to support it. For example, this
now works:

```
-e black[d] @ file://${PROJECT_ROOT}/scripts/packages/black_editable
```

Closes #3941.

Closes #3942.
2024-05-31 21:08:00 +00:00
Charlie Marsh
8c11f99fdf
Discard fragments when parsing unnamed URLs (#3940)
## Summary

Closes https://github.com/astral-sh/uv/issues/3934.
2024-05-31 13:54:15 +00:00
konsti
72b1642232
Move metadata into its own file (#3939)
Move `Metadata`, `MetadataLoweringError` and `ArchiveMetadata` into
their own file `metadata.rs` in `uv-distribution`, moving it out from
`lib.rs`. No functional changes.
2024-05-31 15:24:10 +02:00
konsti
3c074142f5
Re-add lowering unit tests (#3935)
Re-add the lowering unit tests removed in #3904. This also adds a
`stop_discovery_at` feature to avoid running actual workspace discovery.
2024-05-31 12:17:49 +00:00
konsti
9bb0679618
Fix nightly cfg checker warnings (#3932)
Fixes these two warnings on nightly:

```
warning: unexpected `cfg` condition name: `codspeed`
 --> crates/bench/src/lib.rs:5:15
  |
5 |     #[cfg(not(codspeed))]
  |               ^^^^^^^^ help: found config with similar value: `feature = "codspeed"`
  |
  = help: expected names are: `clippy`, `debug_assertions`, `doc`, `docsrs`, `doctest`, `feature`, `miri`, `overflow_checks`, `panic`, `proc_macro`, `relocation_model`, `rustfmt`, `sanitize`, `sanitizer_cfi_generalize_pointers`, `sanitizer_cfi_normalize_integers`, `target_abi`, `target_arch`, `target_endian`, `target_env`, `target_family`, `target_feature`, `target_has_atomic`, `target_has_atomic_equal_alignment`, `target_has_atomic_load_store`, `target_os`, `target_pointer_width`, `target_thread_local`, `target_vendor`, `test`, `ub_checks`, `unix`, and `windows`
  = help: consider using a Cargo feature instead
  = help: or consider adding in `Cargo.toml` the `check-cfg` lint config for the lint:
           [lints.rust]
           unexpected_cfgs = { level = "warn", check-cfg = ['cfg(codspeed)'] }
  = help: or consider adding `println!("cargo::rustc-check-cfg=cfg(codspeed)");` to the top of the `build.rs`
  = note: see <https://doc.rust-lang.org/nightly/rustc/check-cfg/cargo-specifics.html> for more information about checking conditional configuration
  = note: `#[warn(unexpected_cfgs)]` on by default

warning: unexpected `cfg` condition name: `codspeed`
 --> crates/bench/src/lib.rs:8:11
  |
8 |     #[cfg(codspeed)]
  |           ^^^^^^^^ help: found config with similar value: `feature = "codspeed"`
  |
  = help: consider using a Cargo feature instead
  = help: or consider adding in `Cargo.toml` the `check-cfg` lint config for the lint:
           [lints.rust]
           unexpected_cfgs = { level = "warn", check-cfg = ['cfg(codspeed)'] }
  = help: or consider adding `println!("cargo::rustc-check-cfg=cfg(codspeed)");` to the top of the `build.rs`
  = note: see <https://doc.rust-lang.org/nightly/rustc/check-cfg/cargo-specifics.html> for more information about checking conditional configuration
```

```
warning: unexpected `cfg` condition value: `unix`
 --> crates/uv-extract/src/tar.rs:6:16
  |
6 | #[cfg_attr(not(target_os = "unix"), allow(dead_code))]
  |                ^^^^^^^^^^^^^^^^^^
  |
  = note: expected values for `target_os` are: `aix`, `android`, `cuda`, `dragonfly`, `emscripten`, `espidf`, `freebsd`, `fuchsia`, `haiku`, `hermit`, `horizon`, `hurd`, `illumos`, `ios`, `l4re`, `linux`, `macos`, `netbsd`, `none`, `nto`, `openbsd`, `psp`, `redox`, `solaris`, `solid_asp3`, `teeos`, `tvos`, `uefi`, `unknown`, `visionos`, `vita`, `vxworks`, `wasi`, `watchos`, and `windows` and 2 more
  = note: see <https://doc.rust-lang.org/nightly/rustc/check-cfg/cargo-specifics.html> for more information about checking conditional configuration
  = note: requested on the command line with `-W unexpected-cfgs`
```

<!--
Thank you for contributing to uv! To help us out with reviewing, please
consider the following:

- Does this pull request include a summary of the change? (See below.)
- Does this pull request include a descriptive title?
- Does this pull request include references to any relevant issues?
-->
2024-05-31 09:35:52 +00:00
konsti
081f20c53e
Add support for tool.uv into distribution building (#3904)
With the change, we remove the special casing of workspace dependencies
and resolve `tool.uv` for all git and directory distributions. This
gives us support for non-editable workspace dependencies and path
dependencies in other workspaces. It removes a lot of special casing
around workspaces. These changes are the groundwork for supporting
`tool.uv` with dynamic metadata.

The basis for this change is moving `Requirement` from
`distribution-types` to `pypi-types` and the lowering logic from
`uv-requirements` to `uv-distribution`. This changes should be split out
in separate PRs.

I've included an example workspace `albatross-root-workspace2` where
`bird-feeder` depends on `a` from another workspace `ab`. There's a
bunch of failing tests and regressed error messages that still need
fixing. It does fix the audited package count for the workspace tests.
2024-05-31 02:42:03 +00:00
Charlie Marsh
09f55482a0
Remove some unused pub use exports (#3930) 2024-05-30 22:26:52 -04:00
Charlie Marsh
a14fe2f6c7
Avoid filtering preferences by --reinstall (#3929)
## Summary

In general, it's not quite right to filter preferences by `--reinstall`
-- we still want to respect existing versions, we just don't want to
respect _installed_ versions. But now that the installed versions and
preferences are decoupled, we can remove this (`--reinstall` is enforced
on the installed versions via the `Exclusions` struct that we pass to
the resolver).

While I was here, I also cleaned up the lockfile preference code to
better match the structure for `requirements.txt`.
2024-05-30 20:19:08 +00:00
Charlie Marsh
438b5c61d0
Remove installed packages from preferences (#3928)
## Summary

I believe that this is not necessary, as the installer packages are
already considered in `CandidateSelector::get_preferred`.

Firstly, note that we never pass both non-empty installed packages _and_
non-empty preferences (the installer routines pass site packages and no
preferences; the resolver routines pass no site packages but lockfile
preferences).

However, in general, if you look at `CandidateSelector::get_preferred`,
and consider what's changing, we now skip the `if let Some(version) =
preferences.version(package_name)` case for installed packages. But we
then check installed packages within that `if`, and in the `else`. So it
seems like we'll still return them in either case?

The only behavior change is in the case that you have multiple versions
of a package installed. Previously, we'd respect one of them, because
`Preferences` takes the last winner (it's a hash map, so we just replace
the package key with the last version we see); but in installed
packages, we always ignore distributions with multiple versions, since
it's indicative of a broken environment. That's a fine change IMO. We
could change `CandidateSelector::get_preferred` to support this if we
wanted to.
2024-05-30 20:09:31 +00:00
Ibraheem Ahmed
261aa2c70a
Port all git functionality to use git CLI (#3833)
## Summary

We currently rely on libgit2 for most git-related functionality.
However, libgit2 has long-standing performance issues, as well as lags
significantly behind git in terms of new features. For these reasons we
now use the git CLI by default for fetching repositories
(https://github.com/astral-sh/uv/pull/1781). This PR completely drops
libgit2 in favor of the git CLI for all git-related functionality, which
should allow us to use features such as partial clones and sparse
checkouts in the future for performance.

There is also a lot of technical debt in the current git code as it's
mostly taken from Cargo. Switching to the git CLI *vastly* simplifies
the `uv-git` codebase.

Eventually we might want to look into switching to
[`gitoxide`](https://github.com/Byron/gitoxide), but it's currently too
immature for our use case.
2024-05-30 15:28:48 -04:00
Ibraheem Ahmed
85183c1c36
Custom lock-file serialization (#3909)
## Summary

This PR changes the lock-file format to use inline tables for wheels and
source distributions, which currently use separate tables that make the
file harder to follow.

```diff
[[distribution]]
name = "typing-extensions"
version = "4.10.0"
source = "registry+https://pypi.org/simple"

- [distribution.sdist]
- url = "0d26ce356c/typing_extensions-4.10.0.tar.gz"
- hash = "sha256:b0abd7c89e8fb96f98db18d86106ff1d90ab692004eb746cf6eda2682f91b3cb"
- size = 77558
-
- [[distribution.wheel]]
- url = "dc04a3ea60/typing_extensions-4.10.0-py3-none-any.whl"
- hash = "sha256:69b1a937c3a517342112fb4c6df7e72fc39a38e7891a5730ed4985b5214b5475"
- size = 33926

+ sdist = { url = "0d26ce356c/typing_extensions-4.10.0.tar.gz", hash = "sha256:b0abd7c89e8fb96f98db18d86106ff1d90ab692004eb746cf6eda2682f91b3cb", size = 77558 }
+ wheel = [{ url = "dc04a3ea60/typing_extensions-4.10.0-py3-none-any.whl", hash = "sha256:69b1a937c3a517342112fb4c6df7e72fc39a38e7891a5730ed4985b5214b5475", size = 33926 }]
```

The downside is that the inline-tables end up quite long and TOML
doesn't support line breaks in inline tables, yet.

Part of https://github.com/astral-sh/uv/issues/3611.
2024-05-30 19:08:29 +00:00
Andrew Gallant
d3b7d800ea uv-resolver: fix perf regression
We significantly regressed performance in some cases because we were
cloning the resolver state one more time than we needed to. That doesn't
sound like a lot, but in the case where there are no forks, it implies
we were cloning the state for every `get_dependencies` called when we
shouldn't have been cloning it at all.

Avoiding the clone results in somewhat tortured code. This can probably
be refactored by moving bits out to a helper routine, but that also
seemed non-trivial. So we let this suffice for now.
2024-05-30 14:23:14 -04:00
Andrew Gallant
17c043536b uv-resolver: thread markers through the resolver and into the lock file
This addresses the lack of marker support in prior commits.
Specifically, we add them as a new field to `AnnotatedDist`, and from
there, they get added to a `Distribution` in a `Lock`.
2024-05-30 14:23:14 -04:00
Andrew Gallant
f865406ab4 uv-resolver: implement merging of forked resolutions
This commit is a pretty invasive change that implements the merging
of resolutions created by each fork of the resolver.

The main idea here is that each `SolveState` is converted into a
`Resolution` (a new type) and stored on the heap after its fork
completes. When all forks complete, they are all merged into a single
`Resolution`. This `Resolution` is then used to build a `ResolutionGraph`.

Construction of `ResolutionGraph` mostly stays the same (despite the
gnarly diff due to an indent change) with one exception: the code to
extract dependency edges out of PubGrub's state has been moved to
`SolveState::into_resolution`. The idea here is that once a fork
completes, we extract what we need from the PubGrub state and then
throw it away. We store these edges in our own intermediate type which
is then converted into petgraph edges in the `ResolutionGraph`
constructor.

One interesting change we make here is that our edge
data is now a `Version` instead of a `Range<Version>`. I don't think
`Range<Version>` was actually being used anywhere, so this seems okay?
In any case, I think `Version` here is correct because a resolution
corresponds to specific dependencies of each package. Moreover, I didn't
see an easy way to make things work with `Range<Version>`. Notably,
since we no longer have the guarantee that there is only one version of
each package, we need to use `(PackageName, Version)` instead of just
`PackageName` for inverted lookups in `ResolutionGraph::from_state`.

Finally, the main resolver loop itself is changed a bit to track all
forked resolutions and then merge them at the end.

Note that we don't really have any dealings with markers in this commit.
We'll get to that in a subsequent commit.
2024-05-30 14:23:14 -04:00
Andrew Gallant
9e977aa1be uv-resolver: slightly simplify ResolutionGraph::from_state
This changes the constructor to just take an `InMemoryIndex`
directly instead of the constituent parts. No real reason other
than it seems a little simpler.
2024-05-30 14:23:14 -04:00
Andrew Gallant
6f76a66510 uv-resolver: implement basic resolver forking
There are still some TODOs/FIXMEs here, but this makes represents a
chunk of the resolver refactoring to enable forking. We don't do any
merging of resolutions yet, so crucially, this code is broken when no
marker environment is provided. But when a marker environment is
provided, this should behave the same as a non-forking resolver. In
particular, `get_dependencies_forking` is just `get_dependencies`
whenever there's a marker environment.
2024-05-30 14:23:14 -04:00
Andrew Gallant
f5f330627b pep508: make 'simplify_extras' more flexible
This makes it so we can pass any function to determine whether an extra
is always true or not.

For example, `markers.simplify_extras_with(|_| true)` will remove all
extras in any marker expression. This wasn't possible to express
(without knowing all of the marker names) using the old API, but becomes
trivial to express with a predicate function.
2024-05-30 14:23:14 -04:00
Andrew Gallant
a74c2c6792 pep508: add and and or mutators to MarkerTree
While this could be done by callers since the representation
of `MarkerTree` is public, they are just annoying enough to do
that I think it makes sense to provide them on `MarkerTree`
itself.

These could also be improved in the future to do even more
flattening of conjunctions/disjunctions (or perhaps even
more robust simplification). But for now, some basic flattening
is good enough.

These routines will be used to combine marker expressions when
merging forked resolutions.
2024-05-30 14:23:14 -04:00
Charlie Marsh
144566907e
Use lockfile versions as resolution preferences (#3921)
## Summary

Ensures that we avoid upgrading packages unless `--upgrade` or similar
is passed.

For now, the resolver only respects these for registry distributions.

Closes https://github.com/astral-sh/uv/issues/3918.
2024-05-30 17:59:53 +00:00
Charlie Marsh
502e04200d
Remove from_registry_built_dist2 (#3922)
## Summary

I think this was left over from a prior refactor. (The existing
`from_registry_built_dist` was unused.)
2024-05-30 01:57:19 +00:00
Charlie Marsh
a1a5155d34
Move preference behavior out of resolve (#3919)
## Summary

Decoupling this behavior to give more control to clients. No change in
behavior.
2024-05-30 01:16:53 +00:00