Includes the changes from https://github.com/astral-sh/uv/pull/6071 but
takes them way further.
When we have the set of available versions for a package, we can do a
much better job displaying an error.
For example:
```
❯ uv add 'httpx>999,<9999'
× No solution found when resolving dependencies:
╰─▶ Because only the following versions of httpx are available:
httpx<=999
httpx>=9999
and example==0.1.0 depends on httpx>999,<9999, we can conclude that example==0.1.0 cannot be used.
And because only example==0.1.0 is available and you require example, we can conclude that the requirements are unsatisfiable.
```
The resolver has demonstrated that the requested range cannot be used
because there are only versions in ranges _outside_ the requested range.
However, the display of the range of available versions is pretty bad!
We say there are versions of httpx available in ranges that definitely
have no versions available.
With this pull request, the error becomes:
```
❯ uv add 'httpx>999,<9999'
× No solution found when resolving dependencies:
╰─▶ Because only httpx<=1.0.0b0 is available and example depends on httpx>999,<9999, we can conclude that example's
requirements are unsatisfiable.
And because your workspace requires example, we can conclude that your workspace's requirements are unsatisfiable.
```
We achieve this by:
1. Dropping ranges disjoint with the range of available versions, e.g.,
this removes `httpx>=9999`
2. Replacing ranges that capture the _entire_ range of available
versions with the smaller range, e.g., this replaces `httpx<=999` with
`<=1.0.0b0`.
~Note that when we perform (2), we may include an additional bound that
is not relevant, e.g., we include the lower bound of `>=0.6.7`. This is
a bit extraneous, but I don't think it's confusing. We can consider some
advanced logic to avoid that later.~ (edit: I did this, it wasn't hard)
We also improve error messages when there is _only_ one version
available by showing that version instead of a range.
## Summary
Gives the caller control over how messages are reported back to the
user. Also merges the index-location validation into the lock, since
we're already iterating over the packages.
## Summary
This is no longer required since we no longer implement `Eq` on `Lock`.
It will also sometimes be "wrong" as of #6076, since we now apply
different `requires-python` filtering to different parts of the tree
during resolution.
## Summary
Using https://github.com/astral-sh/uv/issues/6064 as a motivating
example: at present, on main, we're not properly propagating the
`Requires-Python` simplifications. In that case, for example, we end up
solving for a branch with `python_version < 3.11`, and a branch `>=
3.11`, even though `Requires-Python` is `>=3.11`. Later, when we get to
the graph, we apply version simplification based on `Requires-Python`,
which causes us to _remove_ the `python_version < 3.11` markers
entirely, leaving us with duplicate dependencies for `pylint`.
This PR instead tries to ensure that we always apply this narrowing to
requirements and forks, so that we don't need to apply the same
simplification when constructing the graph at all.
Closes https://github.com/astral-sh/uv/issues/6064.
Closes#6059.
In particular, I added this as a hack to avoid a kinda of
instability that was caused by our marker code not correctly
detecting markers that were always false. But that has since
been fixed.
Removing this code doesn't change any tests. Arguably it
should be possible to come up with a test that failed with
this hack inserted but succeeded without it. In particular,
with this hack, new forks were being prevented from being
added even when they ought to be added, e.g., when preferences
get updated.
## Summary
This PR changes the definition of `--locked` from:
> Produces the same `Lock`
To:
> Passes `Lock::satisfies`
This is a subtle but important difference. Previous, if
`Lock::satisfies` failed, we would run a resolution, then do
`existing_lock == lock`. If the two weren't equal, and `--locked` was
specified, we'd throw an error.
The equality check is hard to get right. For example, it means that we
can't ship #6076 without changing our marker representation, since the
deserialized lockfile "loses" some of the internal marker state that
gets accumulated during resolution.
The downside of this change is that there could be scenarios in which
`uv lock --locked` fails even though the lockfile would actually work
and the exact TOML would be unchanged. But... I think it's ok if
`--locked` fails after the user modifies something?
Extends https://github.com/astral-sh/uv/pull/6092 to improve resolver
error messages for workspaces that have a single member.
As before, this requires a two-step approach of
1. Traversing the derivation tree and collapsing some members. In this
case, we drop the empty root node in favor of the project.
2. Using special-case formatting for packages. In this case, the
workspace package is referred to with "your project" instead of its
name.
An extension of #6090 that replaces #6066.
In brief,
1. Workspace member names are passed to the resolver for no solution
errors
2. There is a new derivation tree pre-processing step that trims
`NoVersion` incompatibilities for workspace members from the derivation
tree. This avoids showing redundant clauses like `Because only
bird==0.1.0 is available and bird==0.1.0 depends on anyio==4.3.0, we can
conclude that all versions of bird depend on anyio==4.3.0.`. As a minor
note, we use a custom incompatibility kind to mark these
incompatibilities at resolution-time instead of afterwards.
3. Root dependencies on workspace members say `your workspace requires
bird` rather than `you require bird`
4. Workspace member package display omits the version, e.g., `bird`
instead of `bird==0.1.0`
5. Instead of reporting a workspace member as unusable we note that its
requirements cannot be solved, e.g., `bird's requirements are
unsatisfiable` instead of `bird cannot be used`.
6. Instead of saying `your requirements are unsatisfiable` we say `your
workspace's requirements are unsatisfiable` when in a workspace, since
we're not in a "provide direct requirements" paradigm.
As an annoying but minor implementation detail, `PackageRange` now
requires access to the `PubGrubReportFormatter` so it can determine if
it is formatting a workspace member or not. We could probably improve
the abstractions in the future.
As a follow-up, we should additional special casing for "single project"
workspaces to avoid mention of the workspace concept in simple projects.
However, it looks like this will require additional tree manipulations
so I'm going to keep it separate.
## Summary
Historically, in order to "resolve from a lockfile", we've taken the
lockfile, used it to pre-populate the in-memory metadata index, then run
a resolution. If the resolution didn't match our existing resolution, we
re-resolved from scratch.
This was an appealing approach because (in theory) it didn't require any
dedicated logic beyond pre-populating the index. However, it's proven to
be _really_ hard to get right, because it's a stricter requirement than
we need. We just need the current lockfile to _satisfy_ the requirements
provided by the user. We don't actually need a second resolution to
produce the exact same result. And it's not uncommon that this second
resolution differs, because we seed it with preferences, which
fundamentally changes its course. We've worked hard to minimize those
"instabilities", but they're still present.
The approach here is intended to be much simpler. Instead of resolving
from the lockfile, we just check if the current resolution satisfies the
state of the workspace. Specifically, we check if the lockfile (1)
contains all the relevant members, and (2) matches the metadata for all
dependencies, recursively. (We skip registry dependencies, assuming that
they're immutable.)
This may actually be too conservative, since we can have resolutions
that satisfy the requirements, even if the requirements have changed
slightly. But we want to bias towards correctness for now.
My hope is that this scheme will be more performant, simpler, and more
robust.
Closes https://github.com/astral-sh/uv/issues/6063.
## Summary
This was added in https://github.com/astral-sh/uv/pull/5405 but is now
the cause of an instability in `github_wikidata_bot`. Specifically, on
the initial run, we fork in `pydantic==2.8.2`, via:
```
Requires-Dist: typing-extensions>=4.12.2; python_version >= '3.13'
Requires-Dist: typing-extensions>=4.6.1; python_version < '3.13'
```
In the end, we resolve a single version of `typing-extensions`
(`4.12.2`)... But we don't recognize the two resolutions as the "same
graph", because we propagate the fork markers, and so the "edges" have
different markers on them...
In the second run through, when we have the forks in advance, we don't
split on Pydantic... We just try to solve from the root with the current
forks. This is fundamentally different and I fear it will be the cause
of many instabilities. But removing this graph check fixes the proximate
issue.
I don't really understand why this was added since there was no test
coverage in the PR.
## Summary
When constructing the `Resolution`, we only propagated the fork markers
to the package node, but not the extras node. This led to cases in which
an extra could be included unconditionally or otherwise diverge from the
base package version.
Closes https://github.com/astral-sh/uv/issues/6062.
## Summary
Right now, we store the environment markers in a `BTreeSet` -- so
they're sorted, but the sort doesn't really tell us anything. I think we
should instead store them in the order in which we solved. I thought
this might fix an instability (it didn't), but I think it's still good
to ensure we solve in the same order.
I also changed from `Option<Vec>` to just `Vec`, since there was no
distinction between `None` and empty.
## Summary
We retain them if you use `--raw-sources`, but otherwise they're
removed. We still respect them in the subsequent `uv.lock` via an
in-process store.
Closes#6056.
## Summary
Now, if you resolve against a registry, then swap it out for another, we
won't reuse the lockfile. (If you don't provide any registry
configuration, then we won't enforce this, so that `uv lock --index-url
foo` and `uv lock` is stable.)
Closes https://github.com/astral-sh/uv/issues/5920.
This example came up in discussion and it was initially unclear whether
we should try to support it. Specifically, by automatically assuming
that the `datasets < 2.19` dependency had a marker corresponding to the
negation of the conjunction of the other sibling markers for that same
package. But this was deemed, I think, a little too magical.
This in turn implies that whenever there are sibling dependencies with
overlapping marker expressions, their version constraints also need to
be overlapping. Otherwise, for any marker environment that matches both
marker expressions, it would be impossible to select a single version.
The test in this case has this comment:
```
/// If a dependency requests a prerelease version with an overlapping marker expression,
/// we should prefer the prerelease version in both forks.
```
With this setup:
```
let pyproject_toml = context.temp_dir.child("pyproject.toml");
pyproject_toml.write_str(indoc! {r#"
[project]
name = "example"
version = "0.0.0"
dependencies = [
"cffi >= 1.17.0rc1 ; os_name == 'Linux'"
]
requires-python = ">=3.11"
"#})?;
let requirements_in = context.temp_dir.child("requirements.in");
requirements_in.write_str(indoc! {"
cffi
.
"})?;
```
The change in this commit _seems_ more correct that what we had,
although it does seem to contradict the comment. Namely, in the `os_name
!= "Linux"` fork, we don't prefer the pre-release version since the
`cffi >= 1.17.0rc1` bound doesn't apply.
It's not quite clear what to do in this instance.
I believe these are all changes that aren't necessarily
expected, but also seem harmless. Like the order in which
fork markers are written to the lock file. (Although one
wonders if we should fix that once and for all by defining
a complete sort function for forks.)
At a high level, this PR adds a smattering of new tests that
effectively snapshot the output of `uv lock` for a selection of
"ecosystem" projects. That is, real Python projects for which we expect
`uv` to work well with.
The main idea with these tests is to get a better idea of how changes
in `uv` impact the lock files of real world projects. For example,
we're hoping that these tests will help give us data for how #5733
differs from #5887.
This has already revealed some bugs. Namely, re-running `uv lock` for a
second time will produce a different lock file for some projects. So to
prioritize getting the tests added, for those projects, we don't do the
deterministic checking.
