Stream: git-cranelift

Topic: cranelift / PR #1336 Add CI to auto-publish to crates.io ...

GitHub (Jan 10 2020 at 22:34):

alexcrichton opened PR #1336 from auto-publish to master:

This commit updates the publication process for the cranelift suite of
crates. the publish-all.sh script was removed in favor of a new Rust
script located at ci/publish.rs This script is based off the one in
the [wasm-bindgen repository][wbg]. Additionally publication
instructions have been added to CONTRIBUTING.md.

The new process of publication looks like:

• Run rustc ci/publish.rs && ./publish bump
• Commit the change and get it merged (either push it or go through a PR)
• Push a tag with the versions bumped

The CI for the last step, the tag push, will automatically publish the
crates once CI has finished.

[wbg]: https://github.com/rustwasm/wasm-bindgen/blob/master/publish.rs

GitHub (Jan 10 2020 at 22:38):

alexcrichton updated PR #1336 from auto-publish to master:

This commit updates the publication process for the cranelift suite of
crates. the publish-all.sh script was removed in favor of a new Rust
script located at ci/publish.rs This script is based off the one in
the [wasm-bindgen repository][wbg]. Additionally publication
instructions have been added to CONTRIBUTING.md.

The new process of publication looks like:

• Run rustc ci/publish.rs && ./publish bump
• Commit the change and get it merged (either push it or go through a PR)
• Push a tag with the versions bumped

The CI for the last step, the tag push, will automatically publish the
crates once CI has finished.

[wbg]: https://github.com/rustwasm/wasm-bindgen/blob/master/publish.rs

GitHub (Jan 10 2020 at 22:40):

abrown requested bnjbvr for a review on PR #1336.

GitHub (Jan 10 2020 at 22:40):

abrown requested sunfishcode for a review on PR #1336.

GitHub (Jan 10 2020 at 22:40):

abrown requested bnjbvr, and sunfishcode for a review on PR #1336.

GitHub (Jan 10 2020 at 23:30):

sunfishcode submitted PR Review.

GitHub (Jan 10 2020 at 23:30):

sunfishcode submitted PR Review.

GitHub (Jan 10 2020 at 23:30):

sunfishcode created PR Review Comment:

To my knowledge, cranelift-simplejit doesn't do anything special wrt cargo, and doesn't have any unusual dependencies. Can you point to what it is about this crate that causes problems?

GitHub (Jan 10 2020 at 23:30):

sunfishcode created PR Review Comment:

This is unlikely, but could you append and prepend " to the version string, so that 1.1.1 doesn't match 1.1.11?

GitHub (Jan 10 2020 at 23:30):

sunfishcode created PR Review Comment:

Do we need to do anything special to tell it to only publish if this is the master branch, and not some random feature branch?

GitHub (Jan 10 2020 at 23:30):

sunfishcode created PR Review Comment:

Would it be difficult to add a publish subcommand to do this part too? Assuming the merge doesn't get a new sha, it seems like we should be able to pick up the sha from the top of the local repo.

GitHub (Jan 13 2020 at 15:32):

alexcrichton submitted PR Review.

GitHub (Jan 13 2020 at 15:32):

alexcrichton created PR Review Comment:

It's sort of a weird part of Cargo, but if Cargo for some crates decides to put a Cargo.lock in the package, meaning it needs to resolve the crate graph. If a crate depends on something not published on crates.io, however, as would be the case in the "dry run" mode which happens on PRs, then this step will fail because Cargo.lock cannot be generated. I hit this locally when I prefixed all the crates with alexcrichton-* and nothing existed on crates.io.

Cargo decides to put Cargo.lock in crates that contains examples or binaries, and simplejit happened to be the only cranelift crate that met that criteria (had an example)

GitHub (Jan 13 2020 at 15:33):

alexcrichton submitted PR Review.

GitHub (Jan 13 2020 at 15:33):

alexcrichton created PR Review Comment:

This'll only actually do things for tags (previous step enables), and only for tags that start with a v (I'm not great on regexes). Do you think it should be limited further though?

GitHub (Jan 13 2020 at 15:34):

alexcrichton submitted PR Review.

GitHub (Jan 13 2020 at 15:34):

alexcrichton created PR Review Comment:

Sure yeah, I'll edit this in

GitHub (Jan 13 2020 at 15:38):

alexcrichton updated PR #1336 from auto-publish to master:

This commit updates the publication process for the cranelift suite of
crates. the publish-all.sh script was removed in favor of a new Rust
script located at ci/publish.rs This script is based off the one in
the [wasm-bindgen repository][wbg]. Additionally publication
instructions have been added to CONTRIBUTING.md.

The new process of publication looks like:

• Run rustc ci/publish.rs && ./publish bump
• Commit the change and get it merged (either push it or go through a PR)
• Push a tag with the versions bumped

The CI for the last step, the tag push, will automatically publish the
crates once CI has finished.

[wbg]: https://github.com/rustwasm/wasm-bindgen/blob/master/publish.rs

GitHub (Jan 13 2020 at 15:42):

alexcrichton updated PR #1336 from auto-publish to master:

This commit updates the publication process for the cranelift suite of
crates. the publish-all.sh script was removed in favor of a new Rust
script located at ci/publish.rs This script is based off the one in
the [wasm-bindgen repository][wbg]. Additionally publication
instructions have been added to CONTRIBUTING.md.

The new process of publication looks like:

• Run rustc ci/publish.rs && ./publish bump
• Commit the change and get it merged (either push it or go through a PR)
• Push a tag with the versions bumped

The CI for the last step, the tag push, will automatically publish the
crates once CI has finished.

[wbg]: https://github.com/rustwasm/wasm-bindgen/blob/master/publish.rs

GitHub (Feb 24 2020 at 11:50):

bnjbvr requested bnjbvr, and sunfishcode for a review on PR #1336.

GitHub (Feb 28 2020 at 23:36):

alexcrichton closed without merge PR #1336.

Last updated: Nov 22 2024 at 16:03 UTC