New repository_ctx.original_name produces "canonical reproducible" warnings

[mbland]

Description of the bug:

Since commit 8bcfb06 introduced repository_ctx.original_name, closing #24467 and #25121, I've noticed the following "canonical reproducible" warnings (formatted for readability) when using the last_green build (at 4c3863e61136a10564f01d920e4f29e142256e7c as of the timestamp below, in Eastern Standard Time):

2025/02/04 10:28:44 Using unreleased version at commit 4c3863e61136a10564f01d920e4f29e142256e7c 

DEBUG: Rule 'rules_java++toolchains+remote_java_tools'
  indicated that a canonical reproducible form can be obtained
  by dropping arguments ["_original_name"]

DEBUG: Rule 'rules_java++toolchains+remote_java_tools_darwin_arm64'
  indicated that a canonical reproducible form can be obtained
  by dropping arguments ["_original_name"]

DEBUG: Rule 'rules_python++python+python_3_11_aarch64-apple-darwin'
  indicated that a canonical reproducible form can be obtained
  by dropping arguments ["_original_name"]

DEBUG: Rule 'rules_java++toolchains+remotejdk11_macos_aarch64'
  indicated that a canonical reproducible form can be obtained
  by dropping arguments ["_original_name"]

DEBUG: Rule 'rules_java++toolchains+remotejdk21_macos_aarch64'
  indicated that a canonical reproducible form can be obtained
  by dropping arguments ["_original_name"]

DEBUG: Rule 'rules_scala++scala_deps+scala_compiler_source_2_12_14'
  indicated that a canonical reproducible form can be obtained
  by dropping arguments ["_original_name"]

All of these except the last are followed by stanzas similar to:

DEBUG: Repository rules_java++toolchains+remote_java_tools instantiated at:
  <builtin>: in <toplevel>
Repository rule http_archive defined at:
  .../external/bazel_tools/tools/build_defs/repo/http.bzl:392:31: in <toplevel>

None of the following Bazel versions produce this error:

• 6.5.0
• 7.5.0
• 8.0.1
• rolling 9.0.0-pre.20250121.1 (2025-01-30)

Which category does this issue belong to?

External Dependency

What's the simplest, easiest way to reproduce this bug? Please provide a minimal example if possible.

Note that, as I write this, mbland/rules_scala@21e4be9 is the HEAD of the bzlmod-bazel-8 branch of my rules_scala fork. This branch contains my work on bazelbuild/rules_scala#1482 (Bzlmod compatibility) and bazelbuild/rules_scala#1687 (Bazel 8 compatibility).

git clone https://github.com/mbland/rules_scala.git
cd rules_scala
git checkout 21e4be9557645885ba2aabab7922c457ffdf03ee
cd dt_patches/test_dt_patches_user_srcjar
USE_BAZEL_VERSION=4c3863e61136a10564f01d920e4f29e142256e7c bazel clean --expunge  # if not the first build
USE_BAZEL_VERSION=4c3863e61136a10564f01d920e4f29e142256e7c bazel build --repo_env=SCALA_VERSION=2.12.14 //...

Which operating system are you running Bazel on?

macOS Sequoia 15.3 (24D60)

What is the output of bazel info release?

development version

If bazel info release returns development version or (@non-git), tell us how you built Bazel.

Using the last_green build from Bazelisk

What's the output of git remote get-url origin; git rev-parse HEAD ?

4c3863e61136a10564f01d920e4f29e142256e7c

If this is a regression, please try to identify the Bazel commit where the bug was introduced with bazelisk --bisect.

8bcfb06

Have you found anything relevant by searching the web?

It seems a similar problem surfaced in 2020:

• Warning: Rule '<every remote repository rule>' indicated that a canonical reproducible form can be obtained by dropping arguments ["generator_name", "generator_function"] #11040 (comment)
• In repo rules, don't warn about generator_* attributes being non-canonical #11113

Based on the above, it seems the fix may be to add original_name to RepositoryResolvedEvent.IGNORED_ATTRIBUTE_NAMES at:

bazel/src/main/java/com/google/devtools/build/lib/bazel/repository/RepositoryResolvedEvent.java

Lines 239 to 244 in 544887e

	/**
	* Attributes that may be defined on a repository rule without affecting its canonical
	* representation. These may be created implicitly by Bazel.
	*/
	private static final ImmutableSet<String> IGNORED_ATTRIBUTE_NAMES =
	ImmutableSet.of("generator_name", "generator_function", "generator_location");

Any other information, logs, or outputs that you want to share?

In the bzlmod-bazel-8 branch of my rules_scala fork, which contains my work on bazelbuild/rules_scala#1482 and bazelbuild/rules_scala#1687, this causes several dt_patches/dt_patch_test.sh test cases to fail.

The failing test_compiler_srcjar cases, which check for ! grep "canonical reproducible", appear towards the end (formatting slightly to split DEBUG: messages on separate lines):

# Assuming `$PWD` is still `dt_patches/test_dt_patches_user_srcjar`
# after the repro steps above:
$ cd ..

# `$PWD` should now be `dt_patches/`
$ USE_BAZEL_VERSION=4c3863e61136a10564f01d920e4f29e142256e7c ./dt_patch_test.sh

[ ...snip... ]

Starting local Bazel server (no_version) and connecting to it...
INFO: Starting clean (this may take a while). Use --async if the clean takes more than several minutes.
running test test_compiler_srcjar 2.12.14
DEBUG: Rule 'rules_java++toolchains+remote_java_tools_darwin_arm64' indicated that a canonical reproducible form can be obtained by dropping arguments ["_original_name"]
DEBUG: Rule 'rules_java++toolchains+remote_java_tools' indicated that a canonical reproducible form can be obtained by dropping arguments ["_original_name"]
DEBUG: Rule 'rules_python++python+python_3_11_aarch64-apple-darwin' indicated that a canonical reproducible form can be obtained by dropping arguments ["_original_name"]
DEBUG: Rule 'rules_java++toolchains+remotejdk11_macos_aarch64' indicated that a canonical reproducible form can be obtained by dropping arguments ["_original_name"]
DEBUG: Rule 'rules_java++toolchains+remotejdk21_macos_aarch64' indicated that a canonical reproducible form can be obtained by dropping arguments ["_original_name"]
DEBUG: Rule 'rules_scala++scala_deps+scala_compiler_source_2_12_14' indicated that a canonical reproducible form can be obtained by dropping arguments ["_original_name"]
 Test "test_compiler_srcjar 2.12.14" failed  (43 sec)
running test test_compiler_srcjar 2.12.15
DEBUG: Rule 'rules_scala++scala_deps+scala_compiler_source_2_12_15' indicated that a canonical reproducible form can be obtained by dropping arguments ["_original_name"]
 Test "test_compiler_srcjar 2.12.15" failed  (8 sec)
running test test_compiler_srcjar 2.12.16
DEBUG: Rule '+compiler_user_srcjar_repos+scala_compiler_srcjar' indicated that a canonical reproducible form can be obtained by dropping arguments ["_original_name"]
 Test "test_compiler_srcjar 2.12.16" failed  (7 sec)
running test test_compiler_srcjar_nonhermetic 2.12.17
 Test "test_compiler_srcjar_nonhermetic 2.12.17" successful (8 sec)
running test test_compiler_srcjar_nonhermetic 2.12.18
 Test "test_compiler_srcjar_nonhermetic 2.12.18" successful (8 sec)
running test test_compiler_srcjar_nonhermetic 2.12.19
 Test "test_compiler_srcjar_nonhermetic 2.12.19" successful (8 sec)
running test test_compiler_srcjar_nonhermetic 2.12.20
 Test "test_compiler_srcjar_nonhermetic 2.12.20" successful (10 sec)
running test test_compiler_srcjar_nonhermetic 2.13.11
 Test "test_compiler_srcjar_nonhermetic 2.13.11" successful (7 sec)
running test test_compiler_srcjar_nonhermetic 2.13.12
 Test "test_compiler_srcjar_nonhermetic 2.13.12" successful (8 sec)
running test test_compiler_srcjar_nonhermetic 2.13.14
 Test "test_compiler_srcjar_nonhermetic 2.13.14" successful (9 sec)
running test test_compiler_srcjar_nonhermetic 2.13.15
 Test "test_compiler_srcjar_nonhermetic 2.13.15" successful (8 sec)
running test test_compiler_srcjar 3.1.3
DEBUG: Rule 'rules_scala++scala_deps+scala_compiler_source_3_1_3' indicated that a canonical reproducible form can be obtained by dropping arguments ["_original_name"]
 Test "test_compiler_srcjar 3.1.3" failed  (8 sec)
running test test_compiler_srcjar 3.2.2
DEBUG: Rule 'rules_scala++scala_deps+scala_compiler_source_3_2_2' indicated that a canonical reproducible form can be obtained by dropping arguments ["_original_name"]
 Test "test_compiler_srcjar 3.2.2" failed  (8 sec)
running test test_compiler_srcjar_nonhermetic 3.3.5
 Test "test_compiler_srcjar_nonhermetic 3.3.5" successful (8 sec)
running test test_compiler_srcjar 3.4.3
DEBUG: Rule '+compiler_user_srcjar_repos+scala3_compiler_srcjar' indicated that a canonical reproducible form can be obtained by dropping arguments ["_original_name"]
 Test "test_compiler_srcjar 3.4.3" failed  (9 sec)
running test test_compiler_srcjar_nonhermetic 3.5.2
 Test "test_compiler_srcjar_nonhermetic 3.5.2" successful (9 sec)
running test test_compiler_srcjar_nonhermetic 3.6.2
 Test "test_compiler_srcjar_nonhermetic 3.6.2" successful (10 sec)

[fmeum]

@bazel-io fork 8.1.0

[fmeum]

FYI @Wyverald

[iancha1992]

A fix for this issue has been included in Bazel 8.1.0 RC2. Please test out the release candidate and report any issues as soon as possible.
If you're using Bazelisk, you can point to the latest RC by setting USE_BAZEL_VERSION=8.1.0rc2. Thanks!

[mbland]

Can confirm this issue is resolved by Bazel 8.1.0 RC2, as well as the current last_green release from 5e2dfba. Thanks!