drbobbeaty has quit [Read error: Connection reset by peer]
subbu has quit [Quit: Leaving]
<rebelwarrior[m]>
How do I check whether a java version is M1 or x86? (I'm pretty sure it's M1 btw).
subbu has joined #jruby
subbu has quit [Quit: Leaving]
<headius>
I think there's a flag to the Java command that would work but I can't think of it offhand
<headius>
You want to try using the x86 version for now. I found at least one fixable problem and I know of another that will take some work for m1
drbobbeaty has joined #jruby
<rebelwarrior[m]>
Hey just FYI I've looked every which way I could think to search and haven't found a flag that will tell me if a java is x86 or arm from the command line. I think there might not be a flag.
<headius>
Try using `file` to identify the java command
<headius>
I'll be on the job in a bit
<headius>
ok good morning!
<headius>
$ java -Xinternalversion
<headius>
OpenJDK 64-Bit Server VM (11.0.13+8-LTS) for bsd-aarch64 JRE (Zulu11.52+13-CA) (11.0.13+8-LTS), built on Oct 7 2021 08:52:57 by "zulu_re" with clang Apple LLVM 12.0.0 (clang-1200.0.32.28)
<headius>
rebelwarrior: ^
<headius>
try that
<headius>
like I say, M1 support on JRuby is close but not good enough for most nontrivial Ruby stuff, but x86_64 JDK should work fine
<headius>
I will make full M1 support my xmas project
<headius>
I'm looking at the last couple M1 failures in jnr-posix but I believe they are fcntl varargs issues
subbu has joined #jruby
<headius>
yeah two tests for fcntl DUPFD and one for GETFD... since fcntl is varargs I'm pretty sure these are issues passing values in correctly
<headius>
I will spin a release of this jnr-posix so we can start using it
<headius>
jnr-posix 3.1.14 and jnr-unixsocket 0.38.15 have been released to maven central to pick up the fixed stat struct layout
drbobbeaty has quit [Ping timeout: 256 seconds]
drbobbeaty has joined #jruby
<headius>
another domino falls... ostruct has been patched for JRuby (avoid aliasing caller-accessing methods like instance_eval): https://github.com/jruby/jruby/pull/6963
<puritylake[m]>
headius: is there anything pertinent and within my current knowledge range that needs doing?
<headius>
Perhaps! Maybe we can review failing tests/specs on master and see if there's any Ruby 3.0 features that sound approachable
<puritylake[m]>
Awesome, I’ll be on my dev machine in about 30 mins or maybe an hour
<headius>
ok, all the test runs are on GHA so we can poke around there
<headius>
they're red right now so we can see all the things left to be done
<headius>
looks like I broke master's build with some keyword arg changes, just pushed an update that should get it compiling again
<puritylake[m]>
Ok diving in now
<puritylake[m]>
Never used Github Actions so this is a first for me
<headius>
yeah we have a couple dozen jobs so the ones you want will be test-mri-core and spec-ruby I think
<headius>
those would be running CRuby's tests and ruby/spec
<puritylake[m]>
Ok cool, thanks
<headius>
lunch and stuff, bbiab
<puritylake[m]>
No worries, also I'm not seeing either of those jobs tbh
<puritylake[m]>
Oh is it cause of that error you introduced?
<headius>
bleh... well the workflow is "JRuby CI" but the jobs are failing to build because Maven hasn't propagated my jnr updates yet
<headius>
I would have thought an hour would be plenty of time but sometimes it takes longer
<headius>
might be easier for you to just pull master and try to build yourself, or this will sort itself out eventually
<headius>
I'm going to "re-run all jobs" there and hope it does better
<headius>
the jobs you want are mri-core-* and spec-ruby-*
<headius>
doesn't really matter which one, they just permute JIT settings and so on
<puritylake[m]>
Ok thanks
<puritylake[m]>
Some things have already failed
<headius>
yeah looks like they're still failing, might have cached the missing maven artifact flag
<headius>
I'd give it a try locally... let me know if a clean build fails for you
<puritylake[m]>
Will do
<puritylake[m]>
Bot my favourite and least favourite thing about programming is compiling, good cause I can do whatever while it runs, bad cause it could fail at any time
<headius>
yeah we try to make our builds rock solid but when you depend on remote resources anything can happen
<puritylake[m]>
One of the reasons completely remote work probably won't be permanent
<puritylake[m]>
For programming anyway
<puritylake[m]>
Although I do like the idea of working from home
<puritylake[m]>
Picked a bad time to look for work, getting turned down at every turn or outright ignored. Thought I'd pick up open source to pass the time and pad the resume
<puritylake[m]>
Trying to help out with MRI too, but I'm rather nervous to pick an issue to fix, it feels more daunting than here
<puritylake[m]>
Ok forgot to fetch upstream but when I did it mvnw failed on jruby-base
<headius>
Failed to find jnr-posix etc?
<puritylake[m]>
unixsocket and posix
<headius>
Bleh. Ok maybe look for an older job on master that didn't fail to build. Not sure why those artifacts are taking so long to propagate
<puritylake[m]>
Ya, it's complaining that an update interval needs to elapse or that the updates are forced