<headius[m]> enebo: I reverted the psych update because sometime in the past month it stopped passing its own tests on JRuby
<headius[m]> Filed this to track the issue and moved psych update to a PR: https://github.com/ruby/psych/issues/479
<headius[m]> builds should go green again
travis-ci has joined #jruby
<travis-ci> jruby/jruby (jruby-9.2:fbe13bf by Charles Oliver Nutter): The build was fixed. https://travis-ci.com/jruby/jruby/builds/218180840 [179 min 2 sec]
travis-ci has left #jruby [#jruby]
sagax has quit [Ping timeout: 264 seconds]
RomainManni-Buca has quit [Ping timeout: 240 seconds]
RomainManni-Buca has joined #jruby
ur5us has quit [Ping timeout: 264 seconds]
ur5us has joined #jruby
travis-ci has joined #jruby
<travis-ci> jruby/jruby (master:a6b0a3a by Charles Oliver Nutter): The build was fixed. https://travis-ci.com/jruby/jruby/builds/218181303 [211 min 57 sec]
travis-ci has left #jruby [#jruby]
ur5us_ has joined #jruby
ur5us has quit [Ping timeout: 264 seconds]
ur5us has joined #jruby
ur5us_ has quit [Ping timeout: 246 seconds]
ur5us has quit [Ping timeout: 246 seconds]
codeponpon[m] has quit [Quit: Idle for 30+ days]
justinmcp_ has joined #jruby
justinmcp has quit [*.net *.split]
subbu is now known as subbu|away
subbu|away is now known as subbu
johnphillips3141 has joined #jruby
<headius[m]> hmmm
<jswenson[m]> trying with 9.2.14.0 now, just to check
<headius[m]> well there was a fix for system + chdir
<headius[m]> supposed to improve argument passing... previously it would generate a shell script string and execute it directly but that screwed up quoting
<headius[m]> that is the sh -c you are seeing
<headius[m]> it passes the dir and all elements of the system call as args to a shell script that does cd followed by exec(args) basically
<headius[m]> it is a clean way for us to pass the system args through without stuffing them in a string and messing up quoting etc
<jswenson[m]> Looks like using backticks instead of system might unblock us
<jswenson[m]> and 9.2.14.0 seems to work as before
<headius[m]> do you have an example command line that gets stuck?
<jswenson[m]> * Looks like using backticks instead of system might unblock us. Edit: maybe not
<headius[m]> this should have worked fine ☹️
<headius[m]> this is the PR that backported these fixes from master: https://github.com/jruby/jruby/pull/6568
<headius[m]> the primary fix was to get chdir + quoting working right but there are other misc items in there
<headius[m]> that first commit there explains the sh -c change
<headius[m]> bleh
<headius[m]> my sh expert said this is posix and should work right everywhere
<headius[m]> I bet it is the introduction of exec that is throwing something off
<headius[m]> but we also wanted to eliminate the intermediate sh process
<jswenson[m]> this is the command that is launching the hanging process: `sh -c cd '/usr/local/.../looker/helltool/jvm-modules'; /nix/store/29w5rv207rcakiabb0hljnfp3nkcd0mv-gradle-6.7/bin/gradle buildHellToolJar`
<headius[m]> can you test a patched 9.2.15 if I give you a patch? Basically removing exec from that sh command line
<jswenson[m]> I haven’t tried building jruby, but maybe johnphillips31416 has?
<headius[m]> it is pretty simple in any case, I will put together a patch
<headius[m]> I want to know this fixes your issue before I pull in my sh guy to help figure out why
<jswenson[m]> let us know when you have a branch and we’ll try to give it a shot
<headius[m]> I guess this will be no worse than before as far as the intermediate sh process but I really had hoped to eliminate that
<headius[m]> enebo: ^ might need .16 soon after all
<headius[m]> frustrating... this has been running fine in CI for months
<headius[m]> jswenson: there would not be any redirects in that command line right?
<headius[m]> I am seeing some posts about hanging exec when there's redirects
<jswenson[m]> I don’t believe so.
<jswenson[m]> don’t think there are any in the chain
<headius[m]> ok
<headius[m]> well I suspect this patch will fix it but I do not have an explanation for why it hangs
<headius[m]> gonna have to get the expert
<headius[m]> this commit is on master only but I don't think it should change this behavior
<headius[m]> jswenson: if this needs a new release we can do it but I want to know exactly why it is broken
<headius[m]> jswenson: and you should open a bug when you get a chance
<jswenson[m]> still trying to build jruby myself
<jswenson[m]> Will do.
<headius[m]> does `gradle` command have an interactive mode?
<headius[m]> main thing I can think of that might cause a problem would be that exec will inherit tty and subprocess might think it is in an interactive terminal then
<headius[m]> that is why exec with redirects hangs... launches bash in interactive mode and bash is sitting there waiting for input
<headius[m]> not sure we can do a headless exec in posix though
<headius[m]> isn't this stuff fun
<headius[m]> jswenson: I am getting back to my planned tasks for today, do what you can to confirm this fix and get more info on what is launched
<headius[m]> if the command it launches is still running when hung, and it is a JVM process, getting a jstack from it would also be helpful
subbu is now known as subbu|away
victori_ has joined #jruby
victori has quit [Quit: ZNC 1.8.2 - https://znc.in]
Freeky has joined #jruby
liamwhiteGitter[ has quit [Ping timeout: 246 seconds]
MarcinMielyskiGi has quit [Ping timeout: 246 seconds]
sureshbabu[m] has quit [Ping timeout: 246 seconds]
JesseChavezGitte has quit [Ping timeout: 246 seconds]
daveg_lookout[m] has quit [Ping timeout: 240 seconds]
FlorianDoubletGi has quit [Ping timeout: 240 seconds]
lopex[m] has quit [Ping timeout: 265 seconds]
boc_tothefuture[ has quit [Ping timeout: 246 seconds]
OlleJonssonGitte has quit [Ping timeout: 246 seconds]
sdlin[m] has quit [Ping timeout: 246 seconds]
ahorek[m] has quit [Ping timeout: 246 seconds]
headius[m] has quit [Ping timeout: 246 seconds]
subbu has joined #jruby
enebo[m]1 has quit [Ping timeout: 260 seconds]
justinmcp has joined #jruby
GabrielMazetto[m has quit [Ping timeout: 240 seconds]
kai[m]1 has quit [Ping timeout: 268 seconds]
jswenson[m] has quit [Ping timeout: 246 seconds]
justinmcp_ has quit [Write error: Broken pipe]
Liothen has quit [Ping timeout: 264 seconds]
Liothen has joined #jruby
johnphillips3141 has quit [Ping timeout: 258 seconds]
byteit101[m] has quit [Ping timeout: 258 seconds]
BlaneDabneyGitte has quit [Ping timeout: 258 seconds]
jpsikorra[m] has quit [Ping timeout: 258 seconds]
ur5us has joined #jruby
subbu|away has quit [Remote host closed the connection]
Freaky has quit [Remote host closed the connection]
TimGitter[m]1 has quit [Ping timeout: 240 seconds]
kares[m] has quit [Ping timeout: 240 seconds]
rdubya[m] has quit [Ping timeout: 240 seconds]
JulesIvanicGitte has quit [Ping timeout: 268 seconds]
souravgoswami[m] has quit [Ping timeout: 268 seconds]
MattPattersonGit has quit [Ping timeout: 268 seconds]
XavierNoriaGitte has quit [Ping timeout: 268 seconds]
KarolBucekGitter has quit [Ping timeout: 268 seconds]
TimGitter[m] has quit [Ping timeout: 268 seconds]
vortex1111[m] has quit [Ping timeout: 268 seconds]
RomainManni-Buca has quit [Ping timeout: 265 seconds]
CharlesOliverNut has quit [Ping timeout: 265 seconds]
UweKuboschGitter has quit [Ping timeout: 265 seconds]
kalenp[m] has quit [Ping timeout: 265 seconds]
ChrisSeatonGitte has quit [Ping timeout: 265 seconds]
chrisseaton[m] has quit [Ping timeout: 265 seconds]
marcheiligers[m] has quit [Ping timeout: 265 seconds]
ur5us_ has joined #jruby
liamwhiteGitter[ has joined #jruby
MarcinMielyskiGi has joined #jruby
sureshbabu[m] has joined #jruby
lopex[m] has joined #jruby
ur5us has quit [Ping timeout: 264 seconds]
JesseChavezGitte has joined #jruby
headius[m] has joined #jruby
boc_tothefuture[ has joined #jruby
jswenson[m] has joined #jruby
sdlin[m] has joined #jruby
OlleJonssonGitte has joined #jruby
ahorek[m] has joined #jruby
daveg_lookout[m] has joined #jruby
GabrielMazetto[m has joined #jruby
FlorianDoubletGi has joined #jruby
enebo[m]1 has joined #jruby
BlaneDabneyGitte has joined #jruby
johnphillips3141 has joined #jruby
byteit101[m] has joined #jruby
jpsikorra[m] has joined #jruby
CharlesOliverNut has joined #jruby
ChrisSeatonGitte has joined #jruby
kares[m] has joined #jruby
TimGitter[m]1 has joined #jruby
chrisseaton[m] has joined #jruby
UweKuboschGitter has joined #jruby
RomainManni-Buca has joined #jruby
JulesIvanicGitte has joined #jruby
XavierNoriaGitte has joined #jruby
souravgoswami[m] has joined #jruby
rdubya[m] has joined #jruby
vortex1111[m] has joined #jruby
TimGitter[m] has joined #jruby
marcheiligers[m] has joined #jruby
KarolBucekGitter has joined #jruby
kalenp[m] has joined #jruby
MattPattersonGit has joined #jruby
ur5us has joined #jruby
kai[m]1 has joined #jruby
ur5us_ has quit [Ping timeout: 265 seconds]
<jswenson[m]> trying to link my custom build into nix now so should be able to verify the change soon
<headius[m]> oh you guys use nix too
<headius[m]> neat
<jswenson[m]> Yeah, I just recently started using it for development.
<jswenson[m]> But we have a few folks who are driving its adoption here
_whitelogger has joined #jruby
Freeky is now known as Freaky
<jswenson[m]> Ok I believe I have successfully built and run this with the patched version and the process still hangs.
ur5us has quit [Ping timeout: 240 seconds]
<jswenson[m]> Seems to hang in the same place. Notably, running gradle --version (compared to an actual gradle build task) seems to complete without problems.
<jswenson[m]> * Seems to hang in the same place. Notably, running `gradle --version` (compared to an actual gradle build task) seems to complete without problems.
<headius[m]> hmmm I see
<headius[m]> I pasted some examples of a bare `gradle` execution not hanging with that sh form at the command line which seems to support it not being the exec
<headius[m]> can you get a thread dump from the parent process?
<headius[m]> also confirm gradle is still running
<headius[m]> I think you hve done that
<jswenson[m]> gradle daemon is running (launched by an earlier attempt where it failed that time too)
<headius[m]> hmmm
<headius[m]> you said using backquote works around it... what is the format of the backquote command in that case?