Stream: git-wasmtime

Topic: wasmtime / issue #12055 Use cargo publish --workspace ins...


view this post on Zulip Wasmtime GitHub notifications bot (Nov 20 2025 at 15:39):

bjorn3 opened issue #12055:

Feature

See title

Benefit

This would allow getting rid of https://github.com/bytecodealliance/wasmtime/blob/main/scripts/publish.rs and in particular having to explicitly specify which packages to publish and in what order. It should also make publishing faster by publishing multiple crates at once and only waiting for the publish to be propagated when publishing a crate that has a just published crate as dependency.

Implementation

TODO: Do you have an implementation plan, and/or ideas for data structures or
algorithms to use?

Alternatives

TODO: What are the alternative implementation approaches or alternative ways to
solve the problem that this feature would solve? How do these alternatives
compare to this proposal?

view this post on Zulip Wasmtime GitHub notifications bot (Nov 20 2025 at 15:53):

alexcrichton commented on issue #12055:

While I'd like to take advantage of this, there's a few reasons I haven't tried to pull the trigger yet:

I've no doubt that publish --workspace is faster, but it's not really the long poll in our publication process anyway, so I'm not too worried about it personally.

view this post on Zulip Wasmtime GitHub notifications bot (Nov 20 2025 at 16:12):

bjorn3 commented on issue #12055:

We'd need a flag like --idempotent. Inevitably publishes go wrong and being unable to restart a publish is a bit of a deal breaker.

:+1:

such as verifying crates build from packaged sources

That is basically cargo package --workspace without --no-verify, right?

view this post on Zulip Wasmtime GitHub notifications bot (Nov 20 2025 at 20:55):

alexcrichton commented on issue #12055:

Almost yeah, we also test that it builds and at least locally cargo package --workspace doesn't look like it's building the crates, it's just packaging them. For the wasmtime repo cargo package --workspace also looks to fail since it's packaging publish = false packages which otherwise aren't compatible with packaging.

view this post on Zulip Wasmtime GitHub notifications bot (Oct 22 2026 at 21:06):

alexcrichton closed issue #12055:

Feature

See title

Benefit

This would allow getting rid of https://github.com/bytecodealliance/wasmtime/blob/main/scripts/publish.rs and in particular having to explicitly specify which packages to publish and in what order. It should also make publishing faster by publishing multiple crates at once and only waiting for the publish to be propagated when publishing a crate that has a just published crate as dependency.

Implementation

TODO: Do you have an implementation plan, and/or ideas for data structures or
algorithms to use?

Alternatives

TODO: What are the alternative implementation approaches or alternative ways to
solve the problem that this feature would solve? How do these alternatives
compare to this proposal?

view this post on Zulip Wasmtime GitHub notifications bot (Oct 22 2026 at 21:06):

alexcrichton commented on issue #12055:

I'm going to close this as the solution we have is I believe workable for now, but should Cargo work in the future we could certainly move to that


Last updated: Oct 31 2026 at 23:03 UTC