<Freaky> yup
_whitelogger has joined #jruby
Antiarc has quit [Quit: ZNC 1.8.2+deb1 - https://znc.in]
Antiarc has joined #jruby
lopex has quit [Quit: Connection closed for inactivity]
sureshbabu[m] has joined #jruby
<sureshbabu[m]> is there any solution for issue SSL Handshake failure, the environment is Jruby 9.1.5.0, all of sudden it appeared. on googling solutions there for java not for jruby. any clue what was missing and what need to be added to make the functionality working?
ravicious[m] has quit [Quit: Idle for 30+ days]
_whitelogger has joined #jruby
lopex has joined #jruby
<kares[m]> there isn't a general solution for an SSL handshake failure ... any Java solution should work for you but it again depends on your setup
<headius[m]> good morning!
chrisseaton[m] has joined #jruby
<headius[m]> enebo: you think it is worth backporting this to 9.2? https://github.com/jruby/jruby/issues/6565
<enebo[m]> Probably
<headius[m]> it is fixed on master, confirmed by byteit101, but I did not apply it to 9.2
<headius[m]> ok, I will see what it entails
<enebo[m]> My only reservation is not knowing how we really differ everywhere on process launching vs MRI
<headius[m]> the related code could probably be copied straight over
<enebo[m]> Like quoting when doing sh -c or CMD /c
<headius[m]> this fix was only done on the native process launching side also
<headius[m]> yeah well we already differ by having to sh for anything with chdir
<headius[m]> the mrnoname fix at least preserves the original arguments intact now
<enebo[m]> ok so then we would fix it on unixy places and native enabled at that
<headius[m]> it is arguably better than CRuby logic when using sh -c
<enebo[m]> My take is usually how much we can see the boxing of impact if it has a problem
<enebo[m]> Also how much it changes a complicated code base
<enebo[m]> Obviously we are just talking about risk vs benefit
<headius[m]> the risk is probably pretty low... according to mrnoname this form is POSIX sh behavior so it ought to work everywhere
<headius[m]> this is the related PR which does a minor overhaul of the process logic, so that is a bigger risk: https://github.com/jruby/jruby/pull/6226
drbobbeaty has quit [Ping timeout: 256 seconds]
<headius[m]> I think I can pull in just the functional fixes but it would be easier to just pull them all in
<headius[m]> let me see if I can respin the branch as rebased on 9.2 and we can see how we feel about it
drbobbeaty has joined #jruby
<enebo[m]> ok that PR looks large until you realize it is 70% making boolean attributes from a flag
<headius[m]> yeah
<enebo[m]> I cannot easily see if quoting logic changed since it moved
<headius[m]> I know it is hard to review
<headius[m]> lots of commented code and things we will never support finally just deleted
<enebo[m]> but that I suspect would be my only concern
<enebo[m]> if quoting changed the actual changes to processing the text I think would be the main concern
<headius[m]> we are on a path to maintain 9.2 for a while though so having the code match here would be helpful
<enebo[m]> yeah
<enebo[m]> I don't find moving this all back as problematic
<enebo[m]> most of it is just changing one state for another
<headius[m]> this is the main commit for chdir issue
<enebo[m]> There is a tiny risk someone was using those methods but that is so miniscule I am not bothered
<headius[m]> essentially just removes logic that globbed the args into a string and now passes them as argv in this special way
<enebo[m]> ah ok
<headius[m]> a further fix from mrnoname avoids the use of $0 for security reasons
<enebo[m]> makes sense
<enebo[m]> tiny minor comment...that should probably be a literal array create
<enebo[m]> whoop ignote that
<enebo[m]> it can't
<enebo[m]> argc
<enebo[m]> ok
<headius[m]> yeah it is inserting this sh stuff in front of the user args
<enebo[m]> yeah looking in a hurry :)
<headius[m]> my first time using rebase --onto seems to have worked ok
<headius[m]> a few tweaks needed to be green (excludes removed that are only fixed on master, tags not being observed due to differences in mspec config) but it looks good
fidothe has quit [Remote host closed the connection]
fidothe has joined #jruby
<headius[m]> woot... handle_interrupt tweaks have us down to one failure
<headius[m]> guess we were not that far off
fidothe has quit [Remote host closed the connection]
fidothe has joined #jruby