Previously, `simplify_conflict_markers` assumed that it can remove all
conflict set together, when we need to look at each conflict set
individually. Specifically, `(platform_machine == 'x86_64' and extra ==
'extra-5-foo-b') or extra == 'extra-5-foo-a'` can't be reduced
`platform_machine == 'x86_64'` only because it reduces to true when both
conflict extras are activated.
This case applied in https://github.com/astral-sh/uv/issues/14805, where
a jax 0.5.3 version was used for `platform_machine != 'aarch64' or
sys_platform != 'linux'` and the conflict extra `cu128`, but jax 0.7.0
for the conflict extra `cpu`.
Only removing the faulty inference regresses lockfiles to much more
verbose markers. To balance the much more conservative inference, I
added `unify_inference_sets` to simplify cases where all conflict
branches reduce to the same marker.
This still regresses some markers. For example `sys_platform == 'win32'`
regresses to `sys_platform == 'win32' or (extra == 'extra-3-pkg-x1' and
extra == 'extra-3-pkg-x2')` in `extra_inferences`, even through x1 and
x2 conflict and the second conjunction could be simplified away.
Fixes https://github.com/astral-sh/uv/issues/14805
## Summary
Make the use of `Self` consistent. Mostly done by running `cargo clippy
--fix -- -A clippy::all -W clippy::use_self`.
## Test Plan
<!-- How was it tested? -->
No need.
Apply fixes for some `cargo check` and `cargo clippy` lints that are on
in nightly Rust.
The following command now passes, the blanket allows had to many
false-positives:
```
cargo +nightly clippy -- -A clippy::doc_markdown -A mismatched_lifetime_syntaxes -A clippy::explicit_deref_methods
```
`cargo +nightly check -- -A mismatched_lifetime_syntaxes` now passes
without warnings.
Adds `exclude-newer-package = { package = timestamp, ... } ` and
`--exclude-newer-package package=timestamp`. These take precedence over
`exclude-newer` for a given package.
This does need to be serialized to the lockfile, so the revision is
bumped to 3. I tested a previous version and we can read a lockfile with
this information just fine.
Closes https://github.com/astral-sh/uv/issues/14394
`Candidate` has an optional field `prioritized`, which was mostly
redundant with `CandidateDist`. Specifically, it was only `None`, if
`CandidateDist` was `Installed`. This commit removes this duplication.
## Summary
This PR adds derivation chain for another class of resolver failures.
For example, if we encounter a transitive URL dependency, we now tell
the user which package included it, and the full derivation chain:
```
× Failed to resolve dependencies for `foo` (v0.1.0)
╰─▶ Package `flask` was included as a URL dependency. URL dependencies must be
expressed as direct requirements or constraints. Consider adding `flask @
9d4508e893f34853a30fd769c02e9d/flask-3.1.1-py3-none-any.whl`
to your dependencies or constraints file.
help: `foo` (v0.1.0) was included because `baz` (v0.1.0) depends on `foo`
```
Closes#14795.
If a user specifies `-e /path/to/dir` and `/path/to/dir` in a `uv pip
install` command, we want the editable to "win" (rather than erroring
due to conflicting URLs). Unfortunately, this behavior meant that when
you requested a package as editable and non-editable in conflicting
groups, the editable version was _always_ used. This PR modifies the
requisite types to use `Option<bool>` rather than `bool` for the
`editable` field, so we can determine whether a requirement was
explicitly requested as editable, explicitly requested as non-editable,
or not specified (as in the case of `/path/to/dir` in a
`requirements.txt` file). In the latter case, we allow editables to
override the "unspecified" requirement.
If a project includes a path dependency twice, once with `editable =
true` and once without any `editable` annotation, those are now
considered conflicting URLs, and lead to an error, so I've marked this
change as breaking.
Closes https://github.com/astral-sh/uv/issues/14139.
## Summary
Allows `--torch-backend=auto` to detect AMD GPUs. The approach is fairly
well-documented inline, but I opted for `rocm_agent_enumerator` over
(e.g.) `rocminfo` since it seems to be the recommended approach for
scripting:
https://rocm.docs.amd.com/projects/rocminfo/en/latest/how-to/use-rocm-agent-enumerator.html.
Closes https://github.com/astral-sh/uv/issues/14086.
## Test Plan
```
root@rocm-jupyter-gpu-mi300x1-192gb-devcloud-atl1:~# ./uv-linux-libc-11fb582c5c046bae09766ceddd276dcc5bb41218/uv pip install torch --torch-backend=auto
Resolved 11 packages in 251ms
Prepared 2 packages in 6ms
Installed 11 packages in 257ms
+ filelock==3.18.0
+ fsspec==2025.5.1
+ jinja2==3.1.6
+ markupsafe==3.0.2
+ mpmath==1.3.0
+ networkx==3.5
+ pytorch-triton-rocm==3.3.1
+ setuptools==80.9.0
+ sympy==1.14.0
+ torch==2.7.1+rocm6.3
+ typing-extensions==4.14.0
```
---------
Co-authored-by: Zanie Blue <contact@zanie.dev>
Users are not (yet) properly familiar with the concept of universal
resolution and its implication that we need to resolve for all possible
platforms and Python versions. Some projects only target a specific
platform or Python version, and users experience resolution errors due
to failures for other platforms. Indicated by the number of questions we
get about it, `tool.uv.environments` for restricting environments is not
well discoverable.
We add a special hint when resolution failed on a fork disjoint with the
current environment, hinting the user to constrain `requires-python` and
`tool.uv.environments` respectively.
The hint has false positives for cases where the resolution failed on a
different platform, but equally fails on the current platform, in cases
where the non-current fork was tried earlier. Given that conflicts can
be based on `requires-python`, afaik we can't parse whether the current
platform would also be affected from the derivation tree.
Two cases not covered by this are build errors as well as install errors
that need `tool.uv.required-environments`.
Rustfmt introduces a lot of formatting changes in the 2024 edition. To
not break everything all at once, we split out the set of formatting
changes compatible with both the 2021 and 2024 edition by first
formatting with the 2024 style, and then again with the currently used
2021 style.
Notable changes are the formatting of derive macro attributes and lines
with overly long strings and adding trailing semicolons after statements
consistently.
## Summary
If you use `--torch-backend=auto`, we want to avoid selecting (e.g.) a
`+cu124` build of `torch` alongside a `+cu126` build of `torchvision`.
This PR restores the `bogus_redirect` test that was
non-deterministically hanging (reverting #13076).
Mismatched package and distribution names were causing uv to hang prior
to #12917 (which added the `bogus_redirect` test). But with that fix, uv
was only checking for mismatched package names on the main thread (and
not the resolver thread). This allowed for a race condition which would
prevent uv from ever doing the check, triggering the original hang
condition. This PR adds the check to the resolver thread to prevent this
race condition.
When an index performs a bogus redirect or otherwise returns a different
distribution name than expected, uv currently hangs.
In the example case, requesting the simple index page for any package
returns the page for anyio. This mean querying the sniffio version map
returns only anyio entries, and the version maps resolves to an anyio
version. When the resolver makes a query for sniffio and waits for it to
resolve, the main thread finds an anyio and resolves only that in the
wait map, causing the hang.
We fix this by checking the name of the returned distribution against
the name of the requested distribution. For good measure, we add the
same check in `Request::Dist` and `Request::Installed`. For performance
and complexity reasons, we don't perform this check in the version map
itself, but only after a candidate distribution has been selected.
---------
Co-authored-by: Zanie Blue <contact@zanie.dev>
## Summary
This PR modifies the requirement source entities to store a (new)
container struct that wraps `IndexUrl`. This will allow us to store
user-defined metadata alongside `IndexUrl`, and propagate that metadata
throughout resolution.
Specifically, I need to store the "kind" of the index (Simple API vs.
`--find-links`), but I also ran into this problem when I tried to add
support for overriding `Cache-Control` headers on a per-index basis: at
present, we have no way to passing around metadata alongside an
`IndexUrl`.
## Summary
This crate is for standards-compliant types, but this is explicitly a
type that's custom to uv. It's also strange because we kind of want to
reference `IndexUrl` on the registry type, but that's in a crate that
_depends_ on `uv-pypi-types`, which to me is a sign that this is off.
## Summary
This is a prototype that I'm considering shipping under `--preview`,
based on [`light-the-torch`](https://github.com/pmeier/light-the-torch).
`light-the-torch` patches pip to pull PyTorch packages from the PyTorch
indexes automatically. And, in particular, `light-the-torch` will query
the installed CUDA drivers to determine which indexes are compatible
with your system.
This PR implements equivalent behavior under `--torch-backend auto`,
though you can also set `--torch-backend cpu`, etc. for convenience.
When enabled, the registry client will fetch from the appropriate
PyTorch index when it sees a package from the PyTorch ecosystem (and
ignore any other configured indexes, _unless_ the package is explicitly
pinned to a different index).
Right now, this is only implemented in the `uv pip` CLI, since it
doesn't quite fit into the lockfile APIs given that it relies on feature
detection on the currently-running machine.
## Test Plan
On macOS, you can test this with (e.g.):
```shell
UV_TORCH_BACKEND=auto UV_CUDA_DRIVER_VERSION=450.80.2 cargo run \
pip install torch --python-platform linux --python-version 3.12
```
On a GPU-enabled EC2 machine:
```shell
ubuntu@ip-172-31-47-149:~/uv$ UV_TORCH_BACKEND=auto cargo run pip install torch -v
Finished `dev` profile [unoptimized + debuginfo] target(s) in 0.31s
Running `target/debug/uv pip install torch -v`
DEBUG uv 0.6.6 (e95ca063b 2025-03-14)
DEBUG Searching for default Python interpreter in virtual environments
DEBUG Found `cpython-3.13.0-linux-x86_64-gnu` at `/home/ubuntu/uv/.venv/bin/python3` (virtual environment)
DEBUG Using Python 3.13.0 environment at: .venv
DEBUG Acquired lock for `.venv`
DEBUG At least one requirement is not satisfied: torch
warning: The `--torch-backend` setting is experimental and may change without warning. Pass `--preview` to disable this warning.
DEBUG Detected CUDA driver version from `/sys/module/nvidia/version`: 550.144.3
...
```
## Summary
The order here is slightly off... As-is, we fetch the metadata for the
dependency, _then_ insert the URLs and indexes into the fork state -- so
the fetch doesn't take the explicit index or URL into account. This has
mostly been unobserved because we re-fetch anyway in the next request,
but if we do things in the right order (add to fork state, fetch
dependencies, insert dependencies), we can cut down on the fetches.
Closes https://github.com/astral-sh/uv/issues/12056.
Solving spent a chunk of its time just converting resolutions, the left
two blocks:
These blocks are `ResolverOutput::from_state` with 1.3% and
`ForkState::into_resolution` with 4.1% of resolver thread runtime for
apache airflow universal.
We reduce the overhead spent in those functions, to now 1.1% and 2.1% of
resolver time spend in those functions by:
Commit 1: Replace the hash set for the edges with a vec in
`ForkState::into_resolution`. We deduplicate edges anyway when
collecting them, and the hash-and-insert was slow.
Commit 2: Reduce the distribution clonign in
`ResolverOutput::from_state` by using an `Arc`.
The same profile excerpt for the resolver with the branch (note that
there is now an unrelated block between the two we optimized):
Wall times are noisy, but the profiles show those changes as
improvements.
```
$ hyperfine --warmup 2 "./uv-main pip compile --no-progress scripts/requirements/airflow.in --universal" "./uv-branch pip compile --no-progress scripts/requirements/airflow.in --universal"
Benchmark 1: ./uv-main pip compile --no-progress scripts/requirements/airflow.in --universal
Time (mean ± σ): 99.1 ms ± 3.8 ms [User: 111.8 ms, System: 115.5 ms]
Range (min … max): 93.6 ms … 110.4 ms 29 runs
Benchmark 2: ./uv-branch pip compile --no-progress scripts/requirements/airflow.in --universal
Time (mean ± σ): 97.1 ms ± 4.3 ms [User: 114.8 ms, System: 112.0 ms]
Range (min … max): 90.9 ms … 112.4 ms 29 runs
Summary
./uv-branch pip compile --no-progress scripts/requirements/airflow.in --universal ran
1.02 ± 0.06 times faster than ./uv-main pip compile --no-progress scripts/requirements/airflow.in --universal
```
## Summary
This PR revives https://github.com/astral-sh/uv/pull/10017, which might
be viable now that we _don't_ enforce any platforms by default.
The basic idea here is that users can mark certain platforms as required
(empty, by default). When resolving, we ensure that the specified
platforms have wheel coverage, backtracking if not.
For example, to require that we include a version of PyTorch that
supports Intel macOS:
```toml
[project]
name = "project"
version = "0.1.0"
requires-python = ">=3.11"
dependencies = ["torch>1.13"]
[tool.uv]
required-platforms = [
"sys_platform == 'darwin' and platform_machine == 'x86_64'"
]
```
Other than that, the forking is identical to past iterations of this PR.
This would give users a way to resolve the tail of issues in #9711, but
with manual opt-in to supporting specific platforms.
Looks like the set based prioritize tracking from
https://github.com/pubgrub-rs/pubgrub/pull/313 is a slight speedup.
I assume the changed derivation tree in the error snapshot is due to
out-of-sync virtual package priorities, while the main package priority
defining the solution remains stable.
```
$ hyperfine --warmup 2 "./uv-main pip compile --no-progress scripts/requirements/airflow.in --universal" "./uv-branch pip compile --no-progress scripts/requirements/airflow.in --universal"
Benchmark 1: ./uv-main pip compile --no-progress scripts/requirements/airflow.in --universal
Time (mean ± σ): 115.0 ms ± 4.8 ms [User: 131.0 ms, System: 113.6 ms]
Range (min … max): 108.1 ms … 125.8 ms 25 runs
Benchmark 2: ./uv-branch pip compile --no-progress scripts/requirements/airflow.in --universal
Time (mean ± σ): 105.4 ms ± 2.6 ms [User: 118.5 ms, System: 113.5 ms]
Range (min … max): 101.1 ms … 111.9 ms 28 runs
Summary
./uv-branch pip compile --no-progress scripts/requirements/airflow.in --universal ran
1.09 ± 0.05 times faster than ./uv-main pip compile --no-progress scripts/requirements/airflow.in --universal
```
This removes the error that was causing folks problems.
This does result in some snapshot updates that are arguably wrong, or at
least sub-optimal. However, it's actually intended. Because the approach
we're going to take is going to permit the creation of uninstallable
lock files as a side effect. In the future, we will modify this test to
check that, while `uv lock` succeeds, `uv sync` will always fail.
## Summary
This PR reverts https://github.com/astral-sh/uv/pull/10441 and applies a
different fix for https://github.com/astral-sh/uv/issues/10425.
In #10441, I changed prioritization to visit proxies eagerly. I think
this is actually wrong, since it means we prioritize proxy packages
above _everything_ else. And while a proxy only depends on itself, it
does mean we're selecting a _version_ for the proxy package earlier than
anything else. So, if you look at #10828, we end up choosing a version
for `async-timeout` before we choose a version for `langchain`, despite
the latter being a first-party dependency. (`async-timeout` has a marker
on it, so it has a proxy package, so we solve for it first.)
To fix#10425, we instead need to make sure we visit proxies in the
order we see them. I think the virtual tiebreaker for proxies is
reversed? We want to visit the package we see first, first.
So, in short: this reverts #10441, then corrects the ordering for
visiting proxies.
Closes https://github.com/astral-sh/uv/issues/10828.
## Summary
The idea here is to show both (1) an example of a compatible tag and (2)
the tags that were available, whenever we fail to resolve due to an
abscence of matching wheels.
Closes https://github.com/astral-sh/uv/issues/2777.
## Summary
If you have a dependency with a marker, and you add a constraint, it
causes us to _always_ fork, because we represent the constraint as a
second dependency with the marker repeated (and, therefore, we have two
requirements of the same name, both with markers). I don't think we
should fork here -- and in the end it's leading to this undesirable
resolution: #10481.
I tried to change constraints such that we just _reuse_ and augment the
initial requirement, but that has a fairly negative effect on error
messages: #10489. So this fix seems a bit better to me.
Closes https://github.com/astral-sh/uv/issues/10481.
N.B. After fixing #10430, `ArcStr` became the fastest implementation
(and the gains were significantly reduced, down to 1-2%). See:
https://github.com/astral-sh/uv/pull/10453#issuecomment-2583344414.
## Summary
I tried out a variety of small string crates, but `Arc<str>`
outperformed them, giving a ~10% speed-up:
```console
❯ hyperfine "../arcstr lock" "../flexstr lock" "uv lock" "../arc lock" "../compact_str lock" --prepare "rm -f uv.lock" --min-runs 50 --warmup 20
Benchmark 1: ../arcstr lock
Time (mean ± σ): 304.6 ms ± 2.3 ms [User: 302.9 ms, System: 117.8 ms]
Range (min … max): 299.0 ms … 311.3 ms 50 runs
Benchmark 2: ../flexstr lock
Time (mean ± σ): 319.2 ms ± 1.7 ms [User: 317.7 ms, System: 118.2 ms]
Range (min … max): 316.8 ms … 323.3 ms 50 runs
Benchmark 3: uv lock
Time (mean ± σ): 330.6 ms ± 1.5 ms [User: 328.1 ms, System: 139.3 ms]
Range (min … max): 326.6 ms … 334.2 ms 50 runs
Benchmark 4: ../arc lock
Time (mean ± σ): 303.0 ms ± 1.2 ms [User: 301.6 ms, System: 118.4 ms]
Range (min … max): 300.3 ms … 305.3 ms 50 runs
Benchmark 5: ../compact_str lock
Time (mean ± σ): 320.4 ms ± 2.0 ms [User: 318.7 ms, System: 120.8 ms]
Range (min … max): 317.3 ms … 326.7 ms 50 runs
Summary
../arc lock ran
1.01 ± 0.01 times faster than ../arcstr lock
1.05 ± 0.01 times faster than ../flexstr lock
1.06 ± 0.01 times faster than ../compact_str lock
1.09 ± 0.01 times faster than uv lock
```
## Summary
The issue here is that we add `urllib3{python_full_version >= '3.8'}` as
a dependency, then `requests{python_full_version >= '3.8'}`, which adds
`urllib3`, but at that point, we haven't expanded
`urllib3{python_full_version >= '3.8'}`, so we "lose" the singleton
constraint. The solution is to ensure that we visit proxies eagerly, so
that we accumulate constraints as early as possible.
Closes
https://github.com/astral-sh/uv/issues/10425#issuecomment-2580324578.
Ref https://github.com/astral-sh/uv/issues/10344
Not a performance optimization, but the function had become too large.
No logic changes, just code moving around. Looks slightly better when
ignoring whitespace changes.
It's still too complex but i haven't found an apt simplification.
## Summary
Sort of undecided on this. These are already stored as `dyn Reporter` in
each struct, so we're already using dynamic dispatch in that sense. But
all the methods take `impl Reporter`. This is sometimes nice (the
callsites are simpler?), but it also means that in practice, you often
_can't_ pass `None` to these methods that accept `Option<impl
Reporter>`, because Rust can't infer the generic type.
Anyway, this adds more consistency and simplifies the setup by using
`Arc<dyn Reporter>` everywhere.
## Summary
This PR extends #10046 to also handle architectures, which allows us to
correctly include `2.5.1` on the `cu124` index for ARM Linux.
Closes https://github.com/astral-sh/uv/issues/9655.
konsti
adamnemecek
Zanie Blue
Charlie Marsh
John Mumm
Andrew Gallant