Instead of using junctions, we can just write files that contain (as the
file contents) the target path. This requires a little more finesse in
that, as readers, we need to know where to expect these. But it also
means we get to avoid junctions, which have led to a variety of
confusing behaviors. Further, `replace_symlink` should now be on atomic
on Windows.
Closes#11263.
## Summary
We can retain the small-size advantage of our new tags by moving the
"unknown tag" case into `WheelTagLarge`. This ensures that we can still
represent unknown tags, but avoid paying the cost for them.
## Summary
I don't think this had an impact in practice, but it is "wrong" to omit
these. Confirmed that the cache (for example) now includes the build tag
(as in, `mkl_fft-1.3.8-72-cp310-cp310-manylinux2014_x86_64`).
## Summary
This PR extends the thinking in #10525 to platform tags, and then uses
the structured tag enums everywhere, rather than passing around strings.
I think this is a big improvement! It means we're no longer doing ad hoc
tag parsing all over the place.
## Summary
I need to be able to do non-lexicographic comparisons between tags
(e.g., so I can sort `cp313` as greater than `cp39`). It ended up being
easiest to just create structured types for all the tags we support,
with `FromStr` and `Display` implementations.
We don't currently store these in `Tags` or in `WheelFilename`. We may
want to, since they're really small (and `Copy`), but I need to
benchmark to determine whether parsing these in `WheelFilename` is
prohibitively slow.
When building a wheel from a source distribution or both a source
distribution and a wheel, the versions in their filenames must be the
same.
By inspecting the filenames, we also assert that the filenames from the
build a valid (we don't enforce normalization though, just that uv can
parse them).
Note that we're not yet checking that also the `pyproject.toml` version,
if declared, and METADATA version matches.
## Summary
These were moved as part of a broader refactor to create a single
integration test module. That "single integration test module" did
indeed have a big impact on compile times, which is great! But we aren't
seeing any benefit from moving these tests into their own files (despite
the claim in [this blog
post](https://matklad.github.io/2021/02/27/delete-cargo-integration-tests.html),
I see the same compilation pattern regardless of where the tests are
located). Plus, we don't have many of these, and same-file tests is such
a strong Rust convention.
As per
https://matklad.github.io/2021/02/27/delete-cargo-integration-tests.html
Before that, there were 91 separate integration tests binary.
(As discussed on Discord — I've done the `uv` crate, there's still a few
more commits coming before this is mergeable, and I want to see how it
performs in CI and locally).
Charlie Marsh
konsti
Amos Wenger