[Bzlmod] Using `ast-plus` toolchain leads to non reproducibility warning

[FrankPortman]

If you use ast-plus in your toolchain, you will see something like this print out as part of the build:

DEBUG: Rule 'rules_scala++scala_deps+scala_compiler_source_2_13_16' indicated that a canonical reproducible form can be obtained by modifying arguments integrity = "sha256-6SLcL5fVRf3Ag/L88pFesAx2R/N3C5VIneC8nve1Sj4="

You can add the following in your MODULE.bazel to make the error go away:

scala_deps.compiler_srcjar(
    integrity = "sha256-6SLcL5fVRf3Ag/L88pFesAx2R/N3C5VIneC8nve1Sj4=",
    url = "https://repo1.maven.org/maven2/org/scala-lang/scala-compiler/%s/scala-compiler-%s-sources.jar" % (scala_version, scala_version),
    version = scala_version,
)

Ideally, this is not required. But if it is the correct/idiomatic solution, we should better document it.

[FrankPortman]

Slack context: https://bazelbuild.slack.com/archives/CA31HN1T3/p1748006814765239

[mbland]

@simuons Please feel free to assign this to me. This line is the source of the warning:

https://github.com/bazel-contrib/rules_scala/blob/v7.0.0/scala/private/macros/scala_repositories.bzl#L90

It'll take a little work to come up with a way to inject the proper integrity values, but it shouldn't be too bad. I'm thinking it'll be something along the lines of how scripts/update_protoc_integrity.py generates checksums for the protoc binaries.

[mbland]

My fix-1743-scala-compiler-source-warning branch solves this now. I confirmed this by hand by running bazel clean; bazel build //src/... before and after applying the commits on that branch and noticing the output.

Tomorrow I'll clean up the commits, write a proper test, and file a pull request.

[mbland]

@FrankPortman Just filed #1749, which reproduces the problem with a new test and fixes it.

I'm also interested in your thoughts on how I documented how to use scala_deps.compiler_srcjar by including it in the error message from scala/private/macros/scala_repositories.bzl. I'm thinking that might be more helpful in the moment than telling people to go find a doc in the rules_scala repo.

[FrankPortman]

I think the error looks good! I can only dream of always having that much information right in front of me when things fail. Thank you.