Github user jvanzyl commented on the pull request:
https://github.com/apache/maven/pull/34#issuecomment-68680709
Tibor,
Empirically on the daily basis I work on a 900 module build using 8 threads
with 1500+ developers and there aren't many concurrency issues. I'm not keen on
incorporating this patch without a demonstration there is something wrong in
the core. Igor and I took a look at creating a harness to test multithreaded
builds but it's hard. But I think that needs to be done first to show that
something is actually wrong. There may very well be but as I said empirically I
haven't actually seen anything.
I would actually rather go the other way and create a patch that makes
everything immutable and then watch a bunch of the ITs fail and then work
toward making them work, or removing ITs that expect to mutate state. I think
this will be far more informative and provide a path to cleaning up the core.
If I had witnessed all sorts of problems with concurrency issues I might be
inclined to accept the patch but I don't think it particularly helps. Are you
seeing lots of concurrency issues? If the code that clones the project per
build thread is removed, yes, things don't work very well.
I'm happy to discuss more. Maybe pop into IRC, but as-is I'm not inclined
to accept this. Also have you run the Maven ITs with your patch?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
