ur5us has quit [Ping timeout: 252 seconds]
ur5us has joined #jruby
ur5us has quit [Ping timeout: 252 seconds]
<edipofederle[m]> headius: I`m debugging the `rb_export_method` C function, could you give me some tips on how to print the method name?, I would like to inspect it to see which method is currently begin handled
<edipofederle[m]> * headius: Im debugging the `rb_export_method` C function, could you give me some tips on how to print the method name?, I would like to inspect it to see which method is currently begin handled
lopex[m] has quit [Quit: Reconnecting]
lopex[m] has joined #jruby
nirvdrum has joined #jruby
<basshelal[m]> headius: I just posted a new PR. I updated to JUnit5, fixed that git branch problem, turns out was super easy so merging will only add a couple of commits instead of 30+ or whatever. Let me know if you need any clarifications, I've learned from my mistakes and tried to keep things easy to review
<nirvdrum> Howdy. I thought the channel was just dead before realizing everyone moved off FreeNode.
worldnamer[m] has joined #jruby
<worldnamer[m]> Hello!
<nirvdrum> Hi.
<worldnamer[m]> I'm having an issue with JRuby 9.2.19.0 and was wondering if this is the place to go to find out if that issue is legitimate, and what I should do about it.
<edipofederle[m]> worldnamer: hi, maybe you can share the details at https://github.com/jruby/jruby/issues ?
<worldnamer[m]> I'd like to make sure this is a real issue first?
<worldnamer[m]> It's entirely plausible this is just something I'm doing wrong.
<edipofederle[m]> please, share with us, I will try help you.
<worldnamer[m]> I'm running rvm 1.29.12 and jruby 9.2.19.0 on OSX 10.15.7.
<worldnamer[m]> In IRB:
<worldnamer[m]> jruby-9.2.19.0 :002 > return_value, captured_err, status = Open3.capture3('git --version')
<worldnamer[m]> git version 2.32.0
<worldnamer[m]> => ["", "", #<Process::Status: pid 98101 exit 0>]
<worldnamer[m]> I would expect that capture3 would return a non-empty string for the first string.
<worldnamer[m]> But I know that capture3 has been wonky on specific versions, and maybe that's where I'm getting hung up?
<worldnamer[m]> If you think this is something I should report on github, I'm happy to file an issue! Just didn't want to jump the gun without confirming. It can be really tricky with all the moving parts to get this right.
<worldnamer[m]> I also know I have a pending OS update but I'm traveling and was hoping to avoid updating my OS while on the road.
<nirvdrum> worldnamer[m]: If it helps any, I just tried with JRuby 9.2.19.0 and got the following output: ["git version 2.31.1\n", "", #<Process::Status: pid 2670642 exit 0>]
<worldnamer[m]> <nirvdrum "worldnamer: If it helps any, I j"> Thank you, that does suggest I have something odd going on. Are you on OSX?
<nirvdrum> No. This was Ubuntu 21.04.
<worldnamer[m]> So this might be an OS-specific issue.
<edipofederle[m]> @wor
<nirvdrum> I get the version string with JRuby 9.2.17.0 on macOS. I don't have 9.2.19.0 on this machine yet.
<edipofederle[m]> * worldnamer: here the output on mac osx: `["git version 2.21.1 (Apple Git-122.3)\n", "", #<Process::Status: pid 19465 exit 0>]`
<worldnamer[m]> Huh. Okay. So definitely something wonky on my end. I am at a loss with how to find it though
<worldnamer[m]> That's good to know that at least nominally it's SUPPOSED to work and works for others.
<nirvdrum> I did a quick update to 9.2.19.0 on my macOS machine and it works there as well.
<nirvdrum> Do you see the same thing with MRI?
<worldnamer[m]> Would it be possible my shell is causing an issue? I'm using zsh?
<worldnamer[m]> <nirvdrum "Do you see the same thing with M"> Let me check
<edipofederle[m]> <worldnamer[m] "Would it be possible my shell is"> using zsh as well.
<worldnamer[m]> It looks like MRI doesn't exhibit that, no.
<worldnamer[m]> How odd. I guess I need to do a bit more troubleshooting to find out why JRuby is off on this machine. It's possible RVM did something dumb when it installed.
<worldnamer[m]> I'll try to track it down before I submit an issue. Thank you both, this is the exact sort of sanity check I was hoping to find. :)
<edipofederle[m]> just confirm it works in JRuby installed via RVM as well.
<worldnamer[m]> Well, clearly SOMETHING is the issue :P
<worldnamer[m]> I gotta run though. Thank you again, and I'll check back when I have more information!
<edipofederle[m]> worldnamer: let my check on my Linux machine, 1 sec
<edipofederle[m]> all good on Linux as well.
<worldnamer[m]> Can I ask what your JAVA_OPTS is set to?
<worldnamer[m]> AHA
<worldnamer[m]> THAT WAS IT
<worldnamer[m]> Let me clean that up
<nirvdrum> worldnamer[m]: ?
<worldnamer[m]> Without JAVA_OPTS correctly set, Open3 just doesn't work right.
<worldnamer[m]> Oh, hey, nirvdrum Would you happen to be on Java 8?
<nirvdrum> Java 15 on Linux and 11 on macOS.
<worldnamer[m]> Or I guess edipo.federle but alright. I think it's just down to the install of the JDK on my machine, which was a little rushed.
<worldnamer[m]> I don't know if this rises to the level of a git issue; essentially, when the classes are inaccessible, Open3 silently fails. This wouldn't be that big a problem, except that Bundler is using Open3 when it pulls git projects directly (using the gem... git:'...' syntax.)
<worldnamer[m]> So if you have no JAVA_OPTS and you're using bundler to install custom gems, the whole thing just breaks in a really ugly, difficult to understand way.
<worldnamer[m]> Is it worth filing an issue about that?
<enebo[m]> worldnamer: open an issue on it please. We should be putting all the basic opens already
<worldnamer[m]> <enebo[m] "worldnamer: open an issue on it "> Will do. Thanks for the guidance!
<worldnamer[m]> Can I ask an impertinent culture question?
<enebo[m]> worldnamer: sure
<worldnamer[m]> Why Matrix? Wouldn't it be easier to just... spawn a Discord server?
<enebo[m]> worldnamer: we wanted to use something which was an open protocol and not a proprietary product. Admittedly moving to another server would not practically be much different but we mostly were coming from freenode a couple of years ago when it was being DDOSd
<enebo[m]> So a criteria at the time was something where third party connections were possible. Like, for example, we still bridge to irc so people can continue to use irc here
<worldnamer[m]> Fair enough. IRC will never die, and I can appreciate the desire to use an open protocol.
<enebo[m]> I don't mind discord fwiw either but those were our main reasons at the time and we do not like to revisit this stuff very often since people will remember stuff like "oh they are on matrix"
<worldnamer[m]> Yeah, there's something to be said for setting expectations and sticking with them; Discord is where I live and breathe these days, but like, it's a commercial product, so I can get why someone might not want to sign on for "access to us goes through this random company"
<enebo[m]> worldnamer: my only distinct issue with discord is how often they update the client
<enebo[m]> they make that updating pretty painless but it is constant :)
<worldnamer[m]> Very constant. I've gotten several builds in a week lol
<enebo[m]> They give steam a run for their money
<enebo[m]> lopex: I believe joni does not have this yet?
<enebo[m]> I think either 2.7 or 3.0 added it
<lopex[m]> enebo: it does
<lopex[m]> aka keep
<enebo[m]> lopex: is it a config option?
<lopex[m]> no
<lopex[m]> er, syntax one
<enebo[m]> I never even tried I just happened to notice and realized I had never seen it before
<lopex[m]> if (syntax.op2EscCapitalKKeep()) token.type = TokenType.KEEP;
<enebo[m]> I just meant enabling it
<lopex[m]> OP2_ESC_CAPITAL_K_KEEP is enabled for ruby syntax
<enebo[m]> hahah ok never mind
<enebo[m]> it is working already. I thought this was a newer feature
<lopex[m]> it's newish
<enebo[m]> At least $~ seems to be right
<lopex[m]> with that gpos added lately we should be on pas with onigmo features-wise
<lopex[m]> *par even
<enebo[m]> lopex: you know who Sandy Munro is? Look for Munro ID.4 teardown. Looks at EV drivetrain and motor. Seems like it might be down your alley (I also enjoyed the premliinary teardown of their battery pack
<enebo[m]> lopex: For 3.0 is there data changes?
<lopex[m]> I'd have to check
<nirvdrum> lopex[m]: I'm looking a bit at joni today. Have you ever tried caching Matcher instances at a call site?
<lopex[m]> nirvdrum: no
<nirvdrum> Is there any internal state that would get messed up after running a match?
<lopex[m]> yeah, it holds all matching state
<lopex[m]> stack, groups etc
<nirvdrum> Okay. I was afraid of that.
<lopex[m]> er
<lopex[m]> no groups there of course
<nirvdrum> I'm just looking at a profile session and org.joni.StackMachine.<init> is a hot spot.
<lopex[m]> probably garbbing the stack might be
<nirvdrum> I'm flirting with some other ideas around short-circuiting regexps that consist only of literal match characters. I'm seeing cases where people do that instead of String#include?.
<lopex[m]> maybe the skip algos are the overhead for short inputs
<lopex[m]> well, they almost certainly are
<nirvdrum> FWIW, I'm seeing JRuby and TruffleRuby are both faster than MRI, so Joni is doing some good stuff. I'm just trying to see if there's anything else I can squeeze out of there.
<nirvdrum> (In case you didn't hear, I'm back to working on TruffleRuby. This time at Shopify.)
<lopex[m]> nirvdrum: there;s this too https://github.com/jruby/joni/issues/43
<lopex[m]> nirvdrum: also, you might try to compare search vs match
<nirvdrum> Interesting. Thanks for the pointers.
<lopex[m]> enebo: there are jcodings test failures after updating the data
ur5us has joined #jruby
ur5us has quit [Read error: Connection reset by peer]
nirvdrum has quit [Ping timeout: 240 seconds]