BZK[m] has joined #jruby
<headius> Psych 5.1 is released and the PR is updated
<headius> oops, forgot to update psych tests in that PR
<headius> ok I update the tests and the PR is running another cycle, but it can be merged any time; I confirmed they pass locally and they run in CI for psych
<headius> so I think we're basically ready
<enebo[m]> nice
<enebo[m]> There are 3 failures on new psych in test/mri but all related to safe load
<enebo[m]> err 5 failures 4 seem related around safe load and 1 is a parse error itself
<headius> yeah those are all fixed by the updated tests
<headius> yup, test:mri:stdlib is green now
<enebo[m]> ok so we were out of date on some tests and you sync'd mri?
<headius> yeah the psych tests we had were whatever 3.1 used so they're not up to date with the psych repo or MRI master
<enebo[m]> ah ok
<enebo[m]> once that is merged I can build bits and test
<headius> so there is some risk here with loads getting safety treatment but we have to move forward to fix these psych issues
<headius> and the safety stuff in psych has been around for a while... hopefully code that uses unsafe features has already been updated
<headius> ah and I guess some of them were already in the psych we shipped because there was that bug report
<enebo[m]> yeah. we have to move forward on this for sure
<headius> so maybe not as much risk
<enebo[m]> 9.4.1 is not a place where I am so upset with risk
<headius> true
<headius> psych 5.1 merged and the last issues closed
<enebo[m]> I don't want to break stuff either obviously but if there was a better time for this update it only would have been .0
<headius> enebo: we could throw this in with zero risk if you haven't made bits yet: https://github.com/jruby/jruby/issues/7628
<headius> I did not notice it until now
<enebo[m]> I did already upload and close but we can solve that and I will just do it again
<enebo[m]> headius: ^
<headius> ok I've got it done locally
<headius> pushed to master
<headius> also I apparently missed pom.xml changes for psych so that's in as well
<enebo[m]> re-upping
<headius> I just pushed this PR to investigate the other bug reported overnight (rspec ArgumentError on 9.4): https://github.com/jruby/jruby/pull/7631
<headius> we run have been running rspec 3.10 in CI so this may be a new breakage...will have to investigate for 9.4.2
<headius> I'll put together some release notes
<headius> enebo: you have that issue list generated?
<enebo[m]> nope
<enebo[m]> Let me quick do that
<headius> this is marked for 9.4.1 but was not merged: https://github.com/jruby/jruby/pull/7507
<enebo[m]> I also make sure we have nothing closed an untargetted
<enebo[m]> Marked it invalid just now
<headius> big release
<enebo[m]> we did not go the fast turnaround but we did get a lot more people trying it
<enebo[m]> I don't like to see duplicate issues but we may end up stable in a lower number
<enebo[m]> If you consider some of the 9 series it took 15+ releases
<headius> enebo: another one: https://github.com/jruby/jruby/pull/7594
<enebo[m]> Anything I happened to delete without merging can be invalid/duplicate
<headius> ok
<enebo[m]> Admittedly I do not expect to see another one but I was marking against what I was working towards and I know I didn't unmark failures
<enebo[m]> windows seems good
<enebo[m]> I love to see rails 7 just work on windows. Seems really snappy for dev mode too
<headius> yeah it is still impressive to me too
<enebo[m]> hotwire by default means not dealing with webpacker stuff too
<enebo[m]> so I am hoping to see hotwire slowly claw back on JS heavy frontends on Rails
<enebo[m]> Aside...I think I may dualboot my next windows machine so I can try same hardware head to head between Linux and Windows
<enebo[m]> I don't want to mess with repartitioning this box but I am really curious how much worse windows is (which I think will only be from non-native io stuff
<headius> yeah interesting
<headius> IO does still have some challenges on Windows so I'd expect that to be the problem too
<headius> lots of one-off fixes this go-round so not too many release bullets
<enebo[m]> looks good
<enebo[m]> Pass '--add-opens java.base/sun.nio.ch=ALL-UNNAMED --add-opens java.base/java.io=ALL-UNNAMED' to enable.
<enebo[m]> On 17+ jruby-complete on gem list and I imagine everything does this warn
<enebo[m]> Does anyone know a better ps that will group and filter out processes better
<enebo[m]> I tend to ps -fu enebo when trying to kill off soomething and that obviously is all processes owned by me but I need to kill something stuck in another window
<headius> mrnoname had something he used, let me check
<enebo[m]> I am curious how nu shell represents process info since I think as a tree it would be more useful
<enebo[m]> at worst it is tabular but queryable
<enebo[m]> I mean it will be tabular but they may have some builtin filterings
<enebo[m]> like firefox tabs all have one firefox main process parent and we should only see that normally
<enebo[m]> I also realize if I just show processed with a tty it would probably make me totally happy
<enebo[m]> but my brain shrinks every year and ps is the least consistent command in unixy history
<enebo[m]> Testing of bits is done and I see nothing wrong
<enebo[m]> I will push gem and release on maven
<headius> cool, I will start the rest of my bits
<headius> let me know when the site is pushed and I'll do announcements
<enebo[m]> yeah will do . I still have to edit the site so it will be a bit
<enebo[m]> headius: site I will push but not until I see maven artifacts appear
<enebo[m]> which are not there yet
<headius> I need to make this process more automatic one of these days
<headius> yeah waiting for maven
<enebo[m]> lol literally appeared 1s after I wrote that
<enebo[m]> headius: site is live. I am sending emails and will GH release
<headius> ok, I'm working on ruby installers
<headius> enebo: mrnoname uses htop instead of ps
<headius> it does a process tree view
<enebo[m]> ah weirdly I feel like I have asked this before
<headius> he also mentioned pstree
<enebo[m]> htop appears to be a nicer top
<enebo[m]> I guess if I get enough filtering knowledge and options it can pare down to what I want
<headius> he's going to post his htop config
<enebo[m]> pstree is wild
<enebo[m]> GH release is out
<headius> nice
enebo has joined #jruby
ChanServ changed the topic of #jruby to: Get 9.4.1.0! http://jruby.org/ | http://wiki.jruby.org | http://logs.jruby.org/jruby/ | http://bugs.jruby.org | Paste at http://gist.github.com
enebo has left #jruby [#jruby]
<headius> alright, installers and docker are done
<enebo[m]> cool
<headius> title on the release page is 9.4.0.0
<headius> and in the text
<enebo[m]> headius: that rspec report likely is already fixed in 9.4.1.0 since I did fix several issues involving ripper
<headius> fixing
<headius> oh that would be nice
<enebo[m]> ah whoops
<headius> fixed
<enebo[m]> someone also emailed me saying we did not put compat level so I also updated site to include a sentence for that
<enebo[m]> yeah I think someone reported an issue about ripper and I ended up fixing more issues than reported
<enebo[m]> I added a comment to the issue
<enebo[m]> ripper is all new in 9.4 since I combined parsers together
<headius> page build is slow today
<headius> ah right that could be it
<enebo[m]> my fingers are crossed we do not see something significantly wrong but I am not stressed about someone finding a new issue as much as finding a regression
<headius> FB page maintenance is now endless popups trying to get me to use commercial features
<enebo[m]> but we can only test what we have to test with and things seem to be working as well as it was
<enebo[m]> Sounds like amazon music
<enebo[m]> A service I am planning to leave but I need to migrate my mp3vbr library somewhere else
<headius> ok I think that's everything
<headius> I use Spotify now because everything I want is there anyway
<headius> I would suggest Spotify but there's no way to upload your own music
<enebo[m]> soundcloud
<headius> yeah maybe good
<enebo[m]> oh I don't know. I guess if you can upload and keep it private so there are no copyright strikes?
<enebo[m]> It is annoying that what should be trivial things today are more complicated than ever
drbobbeaty has quit [Ping timeout: 255 seconds]
drbobbeaty has joined #jruby
<headius> well I attempted to run a Windows vbox on the new preview of vbox for M1 and it just crashes
<headius> after like 5 minutes of trying to boot