Repository navigation
Switch to trusted publishing - #3517
Conversation
There was a problem hiding this comment.
community review
- The assert_ne changes are unrelated and just to get clippy to pass -> I already opened #3514 for that yesterday
- Does https://github.com/rust-lang/rust-bindgen/blob/3d7ec570ad932708c78a0e66d313bf17b1699854/CONTRIBUTING.md#cutting-a-new-bindgen-release need to be updated?
|
Regarding the second point, I don't think so! Stuff should work as before. This is essentially just an authentication change. |
|
Ok nevermind, the current release workflow is broken, it doesn't work. It has to be implemented in another way, I'll check in with the maintainers. |
This comment has been minimized.
This comment has been minimized.
6f0ebff to
3786215
Compare
3786215 to
e3389f0
Compare
|
This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed. Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers. |
|
Ok, I went back and forth on this a couple of times, the interaction with cargo-dist is tricky. I decided to just run |
|
Updated the docs. Though when doing that, I realized that they were actually out of date, as they described running the publish workflow manually, but that's not how it works right now (before this PR). I wonder if the manual publishing through clicking in GitHub UI actually wouldn't be a better choice for bindgen maintainers. Let's see what @emilio thinks. |
|
Yup. We'll see whether the workflow works as expected, if not, I'll fix it/make adjustments. |
Maybe some change in the CI runner images? |
|
Ok, CI is green again. |
| - name: Publish bindgen (lib) | ||
| run: cargo publish --package bindgen --token ${CARGO_REGISTRY_TOKEN} | ||
| - name: Publish bindgen-cli | ||
| run: cargo publish --package bindgen-cli --token ${CARGO_REGISTRY_TOKEN} |
There was a problem hiding this comment.
can you also invalidate this token? I will remove it from the GitHub Actions secrets. 👍
There was a problem hiding this comment.
Tokens are per user, not per crate, right? So whoever created the token (probably @emilio) would invalidate it.
But the token can't publish bindgen anymore, so even if we can't find the owner of the token, it should be fine.

So that we can track the
bindgencrates inteam.CC @emilio @rust-lang/wg-bindgen
Companion PR: rust-lang/team#2848
After this PR is merged, the
CARGO_REGISTRY_TOKENsecret should be removed from this repository (and invalidated).