Update to the latest semver #2990
Conversation
|
☔ The latest upstream changes (presumably #3023) made this pull request unmergeable. Please resolve the merge conflicts. Note that reviewers usually do not review pull requests until merge conflicts are resolved! Once you resolve the conflicts, you should change the labels applied by bors to indicate that your PR is ready for review. Post this as a comment to change the labels: |
The new release of semver treats `>= 0` the same as the `*` wildcard. It makes sense that these are equivalent, but this did break 3 tests and will technically reject syntax in `Cargo.toml` that was previously accepted.
jtgeibel
force-pushed
the
update/semver
branch
from
8ea5053 to
4819d90
Compare
|
I've rebased and updated the error message. I've also run the following commands to determine crates already using a $ cd ~/.cargo/registry/index/gh.neting.cc-1ecc6299db9ec823/
$ git grep --count '">= 0"' origin/master
origin/master:as/ki/askii:1
origin/master:co/ar/coarsetime:7
origin/master:co/un/count-min-sketch:2
origin/master:cu/rl/curl-sys:31
origin/master:cu/rs/cursive_buffered_backend:4
origin/master:do/h-/doh-proxy:8
origin/master:ju/mp/jumphash:4
origin/master:li/bg/libgit2-sys:38
origin/master:li/bs/libssh2-sys:16
origin/master:na/ts/nats:5
origin/master:pr/iv/privdrop:5
origin/master:vm/-m/vm-memory:2
$ git grep --count '">=0"' origin/master
origin/master:js/on/jsonschema:1
origin/master:vm/-m/vm-memory:5I've checked the index for all of these, and |
|
@bors r+ |
|
📌 Commit 4819d90 has been approved by |
|
☀️ Test successful - checks-actions |
|
Is it possible that this update regressed the rendering of dependency version requirements? #3047 |
|
that does seem likely, yeah. certainly not an intentional change. 😞 |
Save original dependency requirement string in the database when publishing This PR should resolve #3047 by saving the original dependency requirement string in the database when publishing, and returning the original string too when calling the `GET /crates/:crate_id/:version/dependencies` endpoint. Before #2990 we were turning `1.2.3` into `^1.2.3` automatically. After that PR were turning it into `>=1.2.3, <2.0.0`, and now, with this PR, we would keep it at `1.2.3` and it would be up to the API client to prefix the `^`, if necessary. r? `@jtgeibel` /cc `@dtolnay`
This reverts commit 0bcb724. This reverts the revert (rust-lang#3051) of rust-lang#2990. Now that rust-lang#3047 is resolved by rust-lang#3094, it should be fine to bump this again.
This reverts commit 0bcb724. This reverts the revert (rust-lang#3051) of rust-lang#2990. Now that rust-lang#3047 is resolved by
This (partially) reverts commit 0bcb724. This partially reverts the revert (rust-lang#3051) of rust-lang#2990. While we can't bump the version yet, it is worth recovering these changes and documenting the remaining work.
The new release of semver treats
>= 0the same as the*wildcard.It makes sense that these are equivalent, but this did break 3 tests
and will technically reject syntax in
Cargo.tomlthat was previouslyaccepted.
r? @ghost