[
https://issues.apache.org/jira/browse/LUCENE-9868?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17308997#comment-17308997
]
Robert Muir commented on LUCENE-9868:
-------------------------------------
I think hashing could be extremely difficult to debug, if something goes wrong.
It could be frustrating to not know why the hash changed.
I'm thinking typical stuff we've seen before in the past such as non-idempotent
regeneration because of hashmap ordering and stuff like that.
On the other hand: a nightly job that would regenerate and fail if the source
code tree is dirty is easy to implement and debug: {{gradlew regenerate
precommit || git diff}} would work perfectly, and if it failed you would see a
DIFF!
I do agree we need to guard against this, but I'm concerned we are going about
it correctly. I think we might be biased by the large DFA and some desire for
it to happen on every single PR, but I think a smaller step (checking nightly)
would be a significant improvement. We can always back changes out and then fix
them!
And consider theres a lot we currently don't do to test PRs already: in
particular we don't even run the unit tests against them.
> Verify checksums on generated files
> -----------------------------------
>
> Key: LUCENE-9868
> URL: https://issues.apache.org/jira/browse/LUCENE-9868
> Project: Lucene - Core
> Issue Type: Sub-task
> Reporter: Dawid Weiss
> Assignee: Dawid Weiss
> Priority: Minor
>
> This would prevent accidental changes to generated resources/ files.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
