headius has quit [*.net *.split]
greyrat has quit [*.net *.split]
greyrat has joined #jruby
headius has joined #jruby
XavierNoriaGitte has quit [*.net *.split]
lopex[m] has quit [*.net *.split]
CharlesOliverNut has quit [*.net *.split]
OlleJonssonGitte has quit [*.net *.split]
liamwhiteGitter[ has quit [*.net *.split]
MarcinMielyskiGi has quit [*.net *.split]
byteit101[m] has quit [*.net *.split]
rebelwarrior[m] has quit [*.net *.split]
fzakaria[m] has quit [*.net *.split]
FlorianDoubletGi has quit [*.net *.split]
JulesIvanicGitte has quit [*.net *.split]
JulesIvanicGitte has joined #jruby
liamwhiteGitter[ has joined #jruby
fzakaria[m] has joined #jruby
OlleJonssonGitte has joined #jruby
FlorianDoubletGi has joined #jruby
byteit101[m] has joined #jruby
MarcinMielyskiGi has joined #jruby
rebelwarrior[m] has joined #jruby
CharlesOliverNut has joined #jruby
XavierNoriaGitte has joined #jruby
lopex[m] has joined #jruby
fidothe has quit []
fidothe has joined #jruby
<headius> morning folks
<basshelal[m]> Morning 😊
<headius> ruby-core will discuss fileutils being out of sync on Thursday
<headius> not sure if we want to wait for that or just ship with a newer released version
<enebo[m]> hmm
<enebo[m]> I don't really see the harm in shipping the newer version for a .0 especially since it will likely be that version soon for them anyways
<headius> my plan for early this week is to start putting together release notes
<enebo[m]> If there is an issue we will just have to revert or fix upstream and push forward
<headius> yeah I agree
<enebo[m]> I will make some bits this morning just to see how we are doing
NoraHoward[m] has joined #jruby
<headius> enebo: we should do some verification using the snapshot builds of maven artifacts too... at a glance jruby-base looks ok but it is a new artifact and we have not tried to use it for anything ourselves: https://oss.sonatype.org/content/repositories/snapshots/org/jruby/jruby-base/
<headius> baroquebobcat might stop by today with some JI issues he's having (9.2.19)
<headius> in case I step away
<enebo[m]> headius: sonatype is failing because jruby-core now no longer attaches sources
<headius> wat
<headius> they are there for other artifacts though
<headius> oops, not there for core either
<headius> they are there for base
<headius> not there for stdlib
<headius> only base seems to have the javadocs in 9.3
<enebo[m]> well I have not hit a problem with that yet but complete is pretty far down in the process
<enebo[m]> that = docs
<enebo[m]> looks like shaded/pom is for core now and it looks like it only runs shade unless somehow the inherit takes care of more?
<headius> complete and stdlib in the 9.2.20 snapshots don't have javadocs either
<headius> shaded is core now, yes, and the "core" dir is just the base jar
<enebo[m]> this is about sources and not javadocs
<headius> oh oops
<enebo[m]> ok I thought you found a second problem :)
<enebo[m]> This is weird to me since core is a shaded base
<enebo[m]> createSourcesJar: '${create.sources.jar}',
<headius> yeah only base and stdlib should need the sources
<enebo[m]> I see that within shade but I don't really maven
<headius> no sources in 9.2.20 complete jar either
<headius> snapshot
<enebo[m]> complete depends on core so I guess it feels it should have sources
<enebo[m]> but do we have an artifact that depends on complete pushed to sonatype?
<headius> 9.3 base has sources but stdlib does not
<headius> 9.2.20 snap stdlib has sources
<headius> perhaps it is a transitive problem?
<headius> I don't know if it detects that all shaded resources have sources or what
<enebo[m]> so complete sees core is shaded and then asks for sources?
<headius> complete 9.2.19 does have sources in central: https://repo1.maven.org/maven2/org/jruby/jruby-complete/
<enebo[m]> and core definitely did before base was made and it was also shaded before
<headius> something may have gotten tweaked that broke publishing sources in 9.2.20 and that got merged to master
<enebo[m]> I could try 9.2.20 and confirm it is broken
<enebo[m]> That would lead us aware from jruby-base being involved
<headius> very little has changed in this stuff in the past month
<headius> yeah try that
<enebo[m]> but I have never pushed 9.3 as if I was releasing
<enebo[m]> although I guess snapshots should mostly have similar requirements for things like sources
<headius> right, if 20 works then something is off on master
<headius> I don't see anything obvious in maven/ on 9.2 branch changes
<headius> no changes whatsoever in that dir since January
<headius> and the only change on master was switching to the public jakarta-annotation-api dep
<headius> no leads
<enebo[m]> yeah I noticed that
<enebo[m]> actually let me try that again just in case it was my env
<enebo[m]> I had to kill a shell and I might not have been using right stuff
<enebo[m]> ok 9.2 worked
<enebo[m]> jruby-core makes a -sources too
<headius> hmm ok
<enebo[m]> [INFO] --- maven-source-plugin:2.1.2:jar-no-fork (attach-sources) @ jruby-core ---
<enebo[m]> [INFO] No sources in project. Archive not created.
<headius> hmm
<enebo[m]> so share uses jar provided which is jruby-base so it doesn't really have source
<enebo[m]> but when jruby-complete is marked as using jruby-core it expects it will have source jar? Seems wrong to me
<headius> I see the provided scope in complete config includes jruby-core excludes jruby-base
<headius> not sure why it excludes base but perhaps removing that would have an impact
<headius> might need to dig up mkristian otherwise, the new structure is peculiar
<enebo[m]> exclusion "org.jruby:jruby-base"
<headius> right
<enebo[m]> I can remove it and run it to see what happens I guess
<enebo[m]> it seems pretty obvious that the things we use which were shaded should excluded
<headius> yeah
<enebo[m]> but the shade will change consumers too
<enebo[m]> so we really don't want jruby-base here
<enebo[m]> but I am just guessing at what this line means. I will try it
<enebo[m]> It is a weird line in the context of jruby-core but I suppose this it to let Maven know it should not look at this artifact for resolution
<enebo[m]> headius: here is a weird one. I removed the exclusion and re-ran maven. That exclusion is not even being put into pom.xml
<enebo[m]> headius: sent an email to kristian with you cc'd
<headius> He was responding to another issue late last week
<headius> We may need to consider the possibility of changing the build sooner than later
<headius> Something we can maintain better
<enebo[m]> I find maven to be pretty opaque in how to fix problems
<enebo[m]> I can read the log and realize it decided not to build -sources.jar but why jruby-complete needs it as part of release
<enebo[m]> Just don't know why...and this may not really purely be maven but just standards established by sonatype but it is it's own domain which requires some level of expertise
<headius> I don't know the sonatype rules on that... It shades the dependent libraries but does that count as a transitive thing or a new library that aggregates to others and still needs sources?
<headius> When we shaded libraries in core we didn't include the source of those other libraries so I am a bit confused as well
<enebo[m]> jruby-base is not listed in the resulting pom.xml even if you remove that exclusion
<enebo[m]> so I feel like it only lists immediate transitive deps?
<enebo[m]> although does jruby-complete say it is direct trans dep to asm?
<headius> And I assume removing the exclude did not help?
<enebo[m]> I love to just dismiss maven altogether since I have never liked it but I feel like we just lack some knowledge in how this is supposed to work when pushing
<enebo[m]> headius: that line does not do anything to the xml file
<headius> Maybe check on sonatype pages for push requirements? I have some errands to run but back in a bit
<headius> This has to be caused by the new artifact, probably some minor config tweak needed
<enebo[m]> yeah I am guessing worst case there is an empty source jar which would just solve it too
<enebo[m]> boom
<enebo[m]> nope guess not
<headius> Hmm
<enebo[m]> LOL: [INFO] Building jar: /home/enebo/work/release/shaded/target/jruby-core-
<enebo[m]> I just made a source dir and made a single .java file
<enebo[m]> Not a real solution but good enough to see if the rest of this builds fully
<headius> Nice
<enebo[m]> looks like because we are rewriting asm now it is still looking for original asm name?
<headius> What do you mean?
<headius> We have been rewriting it a long time
<enebo[m]> well it is not looking for the shaded package name though
<enebo[m]> oh it is including base and core
<headius> So it should just be core I guess?
<enebo[m]> HAHAH I don't know but I don't know what -- include -- means in package
<headius> Did stdlib build sources?
<enebo[m]> nope no sources to compile
<enebo[m]> need to afk but I wonder if we need some exclusion logic during package not to include both if that is even what is happening
<headius> I should have access to a MacStadium M1 Mac Mini instance soon for JRuby/JNR usage!
<basshelal[m]> That's good news! It should also build for x86 macOS as well I'd imagine?