10:42 <kares[m]> <headius> "did you pick that over to master..." <- nope I did not - was hoping jruby-9.3 would be merged over

15:23 <headius> kares: I'm not sure we are doing that anymore but we could try

15:24 <headius> changes for 9.4 have been getting more invasive

15:39 <kares[m]> okay I assumed it's still done if not than for sure I want to backport some stuff

15:42 <kares[m]> seeing 4-5 relevant conflicts ... let me see if I can resolve those and show a diff the merge would bring (before commit-ing to master)

15:44 <headius> ok

15:46 <headius> I think maybe we just talked about when we'd stop merging, I don't see anything in logs where we decided

15:53 <kares[m]> yeah guess at some point it would make sense but the merge is "still" not that bad ... at least it seems reasonable

15:54 <kares[m]> btw. I still get the annoying:... (full message at <https://libera.ems.host/_matrix/media/v3/download/libera.chat/56bb1d6896a059e8102c7418a86aa7ccdb764f11>)

15:56 <headius> I think you may be a little behind on maven version

15:57 <headius> once enebo and I got on the same recent version it stopped switching around

15:57 <headius> I have mvn 3.8.7 on mac and no switching

16:01 <kares[m]> here's what the merge from jruby-9.3 looks like after resolving some conflicts: https://github.com/jruby/jruby/compare/master...kares:merge-9.3-to-master?expand=1

16:02 <headius> there's a change missing

16:03 <headius> I hade to modify the write2 write3 to all have the same name due to a bug in the populator generation

16:03 <headius> ugh

16:03 <headius> hmmm maybe it only affected master

16:04 <headius> kares: https://github.com/jruby/jruby/pull/7823

16:04 <headius> I did not see failures on 9.3 but this version of the write split should probably be backported

16:05 <headius> I forgot I had put these into a PR to merge that change to master

16:05 <headius> if you finish this merge I can make the additional change on 9.3 and merge forward again

16:06 <kares[m]> okay did not know about that - seems to have build fine locally, that is as far as I checked

16:06 <kares[m]> do you want me to merge this than? I am pretty much done unless you want to redo it differently ...

16:07 <headius> yeah go ahead

16:07 <headius> oh you know what I bet you fixed it

16:08 <headius> I suspected it was due to a different Java version (8 I think) encountering the methods in a different order so it sets up write(arg[]) but then generates a write invoker that can handle many arities

16:08 <headius> your change to make that deterministic probably will fix it too

16:08 <kares[m]> okay pushing the merge to master

16:08 <headius> yeah fingers crossed I guess

16:09 <kares[m]> ✅

16:24 <headius> looks fine

16:24 <headius> I'll kill my PR

16:41 <kares[m]> update to mvn 3.8.8 resolved the issue - will update the wrapper on both jruby-9.3 and master as I tend to use that

16:41 <headius> ok

16:41 <headius> good thought I had forgotten about that wrapper