## Summary
This is what I consider to be the "real" fix for #8072. We now treat
directory and path URLs as separate `ParsedUrl` types and
`RequirementSource` types. This removes a lot of `.is_dir()` forking
within the `ParsedUrl::Path` arms and makes some states impossible
(e.g., you can't have a `.whl` path that is editable). It _also_ fixes
the `direct_url.json` for direct URLs that refer to files. Previously,
we wrote out to these as if they were installed as directories, which is
just wrong.
By splitting `path` into a lockable, relative (or absolute) and an
absolute installable path and by splitting between urls and paths by
dist type, we can store relative paths in the lockfile.
When creating a lockfile, lock the combined dependencies for all
packages in a workspace. This make the lockfile independent of where you
are in the workspace.
Fixes#3983
## Summary
This PR separates "gathering the requirements" from the rest of the
metadata (e.g., version), which isn't required when installing a
package's _dependencies_ (as opposed to installing the package itself).
It thus ensures that we don't need to build a package when a static
`pyproject.toml` is provided in `pip compile`.
Closes https://github.com/astral-sh/uv/issues/4040.
## Summary
Given `install -e dagster`, we need to assume that the user meant
`install -e ./dagster`, even though `install dagster` should _not_ be
treated as `install ./dagster`. I suspect pip will change this in the
future (since `pip install dagster` does _not_ meant `pip install
./dagster`) but for now it's what users expect.
Closes https://github.com/astral-sh/uv/issues/3994.
## 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.
## 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.
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.
## 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`.
## Summary
This PR just ensures that when running `uv lock` (or `uv run`), we lock
with all extras. When we later install, we'll also _install_ with all
extras, but that will be changed in a future PR.
## Summary
There are a few behavior changes in here:
- We now enforce `--require-hashes` for editables, like pip. So if you
use `--require-hashes` with an editable requirement, we'll reject it. I
could change this if it seems off.
- We now treat source tree requirements, editable or not (e.g., both `-e
./black` and `./black`) as if `--refresh` is always enabled. This
doesn't mean that we _always_ rebuild them; but if you pass
`--reinstall`, then yes, we always rebuild them. I think this is an
improvement and is close to how editables work today.
Closes#3844.
Closes#2695.
Add workspace support when using `-r <path>/pyproject.toml` or `-e
<path>` in the pip interface. It is limited to all-editable
static-metadata workspaces, and tests only include a single main
workspace, ignoring path dependencies in another workspace. This can be
considered the MVP for workspace support: You can create a workspace,
you can install from it, but some options and conveniences are still
missing. I'll file follow-up tickets (support in lockfiles, support path
deps in other workspace, #3625)
There is also support in `uv run`, but we need
https://github.com/astral-sh/uv/issues/3700 first to properly support
using different current projects in the bluejay interface, currently the
resolution and therefore the lockfile depends on the current project.
I'd do this change first (it's big enough already), then #3700, and then
add workspace support properly to bluejay.
Fixes#3404
## Summary
We actually _already_ ignore these (preferences only apply to versions,
not URLs), it just happens later on. This PR thus just avoids crashing.
The behavior is unchanged.
Closes#3822.
When parsing requirements from any source, directly parse the url parts
(and reject unsupported urls) instead of parsing url parts at a later
stage. This removes a bunch of error branches and concludes the work
parsing url parts once and passing them around everywhere.
Many usages of the assembled `VerbatimUrl` remain, but these can be
removed incrementally.
Please review commit-by-commit.
## Summary
Closes https://github.com/astral-sh/uv/issues/3715.
## Test Plan
```
❯ echo "/../test" | cargo run pip compile -
error: Couldn't parse requirement in `-` at position 0
Caused by: path could not be normalized: /../test
/../test
^^^^^^^^
❯ echo "-e /../test" | cargo run pip compile -
error: Invalid URL in `-`: `/../test`
Caused by: path could not be normalized: /../test
Caused by: cannot normalize a relative path beyond the base directory
```
I don't really understand why this only happens on windows clippy and
not on linux too, but as usual, boxing the error variant fixes it.
Fixup for #3585
Add minimal support for workspace discovery, only used for determining
paths in the bluejay commands.
We can now discover the workspace structure, namely that the
`pyproject.toml` of a package belongs to a workspace `pyproject.toml`
with members and exclusion. The globbing logic is inspired by cargo. We
don't resolve `workspace = true` metadata declarations yet.
This is split out from workspaces support, which needs editables in the
bluejay commands. It consists mainly of refactorings:
* Move the `editable` module one level up.
* Introduce a `BuiltEditableMetadata` type for `(LocalEditable,
Metadata23, Requirements)`.
* Add editables to `InstalledPackagesProvider` so we can use
`EmptyInstalledPackages` for them.
## Summary
This PR introduces parallelism to the resolver. Specifically, we can
perform PubGrub resolution on a separate thread, while keeping all I/O
on the tokio thread. We already have the infrastructure set up for this
with the channel and `OnceMap`, which makes this change relatively
simple. The big change needed to make this possible is removing the
lifetimes on some of the types that need to be shared between the
resolver and pubgrub thread.
A related PR, https://github.com/astral-sh/uv/pull/1163, found that
adding `yield_now` calls improved throughput. With optimal scheduling we
might be able to get away with everything on the same thread here.
However, in the ideal pipeline with perfect prefetching, the resolution
and prefetching can run completely in parallel without depending on one
another. While this would be very difficult to achieve, even with our
current prefetching pattern we see a consistent performance improvement
from parallelism.
This does also require reverting a few of the changes from
https://github.com/astral-sh/uv/pull/3413, but not all of them. The
sharing is isolated to the resolver task.
## Test Plan
On smaller tasks performance is mixed with ~2% improvements/regressions
on both sides. However, on medium-large resolution tasks we see the
benefits of parallelism, with improvements anywhere from 10-50%.
```
./scripts/requirements/jupyter.in
Benchmark 1: ./target/profiling/baseline (resolve-warm)
Time (mean ± σ): 29.2 ms ± 1.8 ms [User: 20.3 ms, System: 29.8 ms]
Range (min … max): 26.4 ms … 36.0 ms 91 runs
Benchmark 2: ./target/profiling/parallel (resolve-warm)
Time (mean ± σ): 25.5 ms ± 1.0 ms [User: 19.5 ms, System: 25.5 ms]
Range (min … max): 23.6 ms … 27.8 ms 99 runs
Summary
./target/profiling/parallel (resolve-warm) ran
1.15 ± 0.08 times faster than ./target/profiling/baseline (resolve-warm)
```
```
./scripts/requirements/boto3.in
Benchmark 1: ./target/profiling/baseline (resolve-warm)
Time (mean ± σ): 487.1 ms ± 6.2 ms [User: 464.6 ms, System: 61.6 ms]
Range (min … max): 480.0 ms … 497.3 ms 10 runs
Benchmark 2: ./target/profiling/parallel (resolve-warm)
Time (mean ± σ): 430.8 ms ± 9.3 ms [User: 529.0 ms, System: 77.2 ms]
Range (min … max): 417.1 ms … 442.5 ms 10 runs
Summary
./target/profiling/parallel (resolve-warm) ran
1.13 ± 0.03 times faster than ./target/profiling/baseline (resolve-warm)
```
```
./scripts/requirements/airflow.in
Benchmark 1: ./target/profiling/baseline (resolve-warm)
Time (mean ± σ): 478.1 ms ± 18.8 ms [User: 482.6 ms, System: 205.0 ms]
Range (min … max): 454.7 ms … 508.9 ms 10 runs
Benchmark 2: ./target/profiling/parallel (resolve-warm)
Time (mean ± σ): 308.7 ms ± 11.7 ms [User: 428.5 ms, System: 209.5 ms]
Range (min … max): 287.8 ms … 323.1 ms 10 runs
Summary
./target/profiling/parallel (resolve-warm) ran
1.55 ± 0.08 times faster than ./target/profiling/baseline (resolve-warm)
```
## Summary
I think this is overall good change because it explicitly encodes (in
the type system) something that was previously implicit. I'm not a huge
fan of the names here, open to input.
It covers some of https://github.com/astral-sh/uv/issues/3506 but I
don't think it _closes_ it.
## Summary
This PR consolidates the concurrency limits used throughout `uv` and
exposes two limits, `UV_CONCURRENT_DOWNLOADS` and
`UV_CONCURRENT_BUILDS`, as environment variables.
Currently, `uv` has a number of concurrent streams that it buffers using
relatively arbitrary limits for backpressure. However, many of these
limits are conflated. We run a relatively small number of tasks overall
and should start most things as soon as possible. What we really want to
limit are three separate operations:
- File I/O. This is managed by tokio's blocking pool and we should not
really have to worry about it.
- Network I/O.
- Python build processes.
Because the current limits span a broad range of tasks, it's possible
that a limit meant for network I/O is occupied by tasks performing
builds, reading from the file system, or even waiting on a `OnceMap`. We
also don't limit build processes that end up being required to perform a
download. While this may not pose a performance problem because our
limits are relatively high, it does mean that the limits do not do what
we want, making it tricky to expose them to users
(https://github.com/astral-sh/uv/issues/1205,
https://github.com/astral-sh/uv/issues/3311).
After this change, the limits on network I/O and build processes are
centralized and managed by semaphores. All other tasks are unbuffered
(note that these tasks are still bounded, so backpressure should not be
a problem).
## Summary
Ensures that we track the origins for requirements regardless of whether
they come from `pyproject.toml` or `setup.py` or `setup.cfg`.
Closes#3480.
This commit touches a lot of code, but the conceptual change here is
pretty simple: make it so we can run the resolver without providing a
`MarkerEnvironment`. This also indicates that the resolver should run in
universal mode. That is, the effect of a missing marker environment is
that all marker expressions that reference the marker environment are
evaluated to `true`. That is, they are ignored. (The only markers we
evaluate in that context are extras, which are the only markers that
aren't dependent on the environment.)
One interesting change here is that a `Resolver` no longer needs an
`Interpreter`. Previously, it had only been using it to construct a
`PythonRequirement`, by filling in the installed version from the
`Interpreter` state. But we now construct a `PythonRequirement`
explicitly since its `target` Python version should no longer be tied to
the `MarkerEnvironment`. (Currently, the marker environment is mutated
such that its `python_full_version` is derived from multiple sources,
including the CLI, which I found a touch confusing.)
The change in behavior can now be observed through the
`--unstable-uv-lock-file` flag. First, without it:
```
$ cat requirements.in
anyio>=4.3.0 ; sys_platform == "linux"
anyio<4 ; sys_platform == "darwin"
$ cargo run -qp uv -- pip compile -p3.10 requirements.in
anyio==4.3.0
exceptiongroup==1.2.1
# via anyio
idna==3.7
# via anyio
sniffio==1.3.1
# via anyio
typing-extensions==4.11.0
# via anyio
```
And now with it:
```
$ cargo run -qp uv -- pip compile -p3.10 requirements.in --unstable-uv-lock-file
x No solution found when resolving dependencies:
`-> Because you require anyio>=4.3.0 and anyio<4, we can conclude that the requirements are unsatisfiable.
```
This is expected at this point because the marker expressions are being
explicitly ignored, *and* there is no forking done yet to account for
the conflict.
## Summary
Fixes https://github.com/astral-sh/uv/issues/1343. This is kinda a first
draft at the moment, but does at least mostly work locally (barring some
bits of the test suite that seem to not work for me in general).
## Test Plan
Mostly running the existing tests and checking the revised output is
sane
## Outstanding issues
Most of these come down to "AFAIK, the existing tools don't support
these patterns, but `uv` does" and so I'm not sure there's an existing
good answer here! Most of the answers so far are "whatever was easiest
to build"
- [x] ~~Is "-r pyproject.toml" correct? Should it show something else or
get skipped entirely~~ No it wasn't. Fixed in
3044fa8b86
- [ ] If the requirements file is stdin, that just gets skipped. Should
it be recorded?
- [ ] Overrides get shown as "--override<override.txt>". Correct?
- [x] ~~Some of the tests (e.g.
`dependency_excludes_non_contiguous_range_of_compatible_versions`) make
assumptions about the order of package versions being outputted, which
this PR breaks. I'm not sure if the text is fairly arbitrary and can be
replaced or whether the behaviour needs fixing?~~ - fixed by removing
the custom pubgrub PartialEq/Hash
- [ ] Are all the `TrackedFromStr` et al changes needed, or is there an
easier way? I don't think so, I think it's necessary to track these sort
of things fairly comprehensively to make this feature work, and this
sort of invasive change feels necessary, but happy to be proved wrong
there :)
- [x] ~~If you have a requirement coming in from two or more different
requirements files only one turns up. I've got a closed-source example
for this (can go into more detail if needed), mostly consisting of a
complicated set of common deps creating a larger set. It's a rarer case,
but worth considering.~~ 042432b200
- [ ] Doesn't add annotations for `setup.py` yet
- This is pretty hard, as the correct location to insert the path is
`crates/pypi-types/src/metadata.rs`'s `parse_pkg_info`, which as it's
based off a source distribution has entirely thrown away such matters as
"where did this package requirement get built from". Could add "`built
package name`" as a dep, but that's a little odd.
When using `tool.uv.sources`, we warn that requirements have a bound,
i.e. at least a lower version constraint.
When using a library, the symbols you import were introduced in
different versions, creating an implicit lower bound. This warning makes
this explicit. This is crucial to prevent backtracking resolvers from
selecting an ancient versions that is not compatible (or worse, doesn't
build), and a performance optimization on top.
This feature is gated to `tool.uv.sources` (as it should have been to
begin with for #3263/#3443) to not unnecessarily break legacy workflows.
It is also helpful specifically when using a `tool.uv.sources` section
that contains constraints that are not published to pypi, e.g. for
workspace dependencies. We can adjust those later to e.g. not constrain
workspace dependencies with `publish = false`, but i think it's the
right setting to start with.
