knu has quit [Read error: Connection reset by peer]
pawnbox has quit [Ping timeout: 256 seconds]
pawnbox_ has quit [Ping timeout: 256 seconds]
pawnbox has joined #jruby
knu has joined #jruby
pawnbox_ has joined #jruby
pawnbox has quit [Ping timeout: 250 seconds]
blaxter has joined #jruby
zacts has quit [Ping timeout: 240 seconds]
pawnbox_ has quit [Remote host closed the connection]
vtunka has joined #jruby
temporalfox has joined #jruby
zacts has joined #jruby
temporalfox has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
pawnbox has joined #jruby
temporalfox has joined #jruby
blaxter has quit [Quit: foo]
temporalfox has quit [Client Quit]
vtunka has quit [Quit: Leaving]
vtunka has joined #jruby
<claudiuinberlin>
Hello, I would like to ask, there is any change this year, or first half of the next year, that we can use, let say, roda/sinatra/sequel with jruby in truffle mode for api related?
temporalfox has joined #jruby
temporalfox has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
temporalfox has joined #jruby
temporalfox has quit [Client Quit]
<chrisseaton>
claudiuinberlin: I think Roda, Sinatra and Sequel already basically work
<chrisseaton>
claudiuinberlin: what's your database?
<claudiuinberlin>
pg
<claudiuinberlin>
there is list of gems, that are compatible with jruby-truffle?
<chrisseaton>
So we'd need to run that C extension, unless you can run with the pg-pr gem
<claudiuinberlin>
like in pypy community?
temporalfox has joined #jruby
<chrisseaton>
There is, but it isn't public at the moment (it's not a secret, just not publicly visible)
<claudiuinberlin>
aa
<chrisseaton>
We're testing against the top 100 gems in CI
<claudiuinberlin>
there are issues with c ext?
<claudiuinberlin>
I guess, there is a java postgres driver
<chrisseaton>
C extensions are quite ready yet - and yes we're aiming to run MRI's C extensions, rather than JRuby's Java extensions
temporalfox has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
drbobbeaty has joined #jruby
shellac has quit [Quit: Computer has gone to sleep.]
shellac has joined #jruby
pawnbox has quit [Ping timeout: 244 seconds]
<GitHub134>
[jruby] pitr-ch pushed 2 new commits to truffle-head: https://git.io/vXEE5
<GitHub134>
jruby/truffle-head ce64121 Petr Chalupa: [Truffle] fix repeat and sub rope combination
<GitHub134>
jruby/truffle-head dcea10f Petr Chalupa: [Truffle] quiet wget
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
temporalfox has joined #jruby
pawnbox has joined #jruby
pawnbox has quit [Remote host closed the connection]
pilhuhn is now known as pil-afk
<GitHub114>
[jruby] eregon pushed 3 new commits to truffle-head: https://git.io/vXEKN
<GitHub114>
jruby/truffle-head 50bae76 Benoit Daloze: [Truffle] Implement part of Thread::Backtrace::Location#label.
<GitHub114>
jruby/truffle-head ca5e762 Benoit Daloze: [Truffle] Copy over Hash capacities rather than a runtime copy + import.
vtunka has quit [Quit: Leaving]
pawnbox has joined #jruby
temporalfox has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
donV has quit [Quit: donV]
donV has joined #jruby
drbobbeaty has joined #jruby
shellac has quit [Quit: Computer has gone to sleep.]
bbrowning_away is now known as bbrowning
shellac has joined #jruby
<kares_>
claudiuinberlin: why do you specifically want jruby-truffle - for experimenting or does classic feel slow?
vtunka has joined #jruby
<chrisseaton>
claudiuinberlin: if you can share your app with us please do, or just tell us your exact stack or give us any representative code
Specialist has joined #jruby
<nirvdrum>
lopex: Are you around?
tcrawley-away is now known as tcrawley
tcrawley is now known as tcrawley-away
tcrawley-away is now known as tcrawley
temporalfox has joined #jruby
drbobbeaty has quit [Ping timeout: 260 seconds]
pil-afk is now known as pilhuhn
tcrawley is now known as tcrawley-away
temporalfox has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
drbobbeaty has joined #jruby
shellac has quit [Quit: Computer has gone to sleep.]
Specialist has quit [Remote host closed the connection]
olle has joined #jruby
temporalfox has joined #jruby
tcrawley-away is now known as tcrawley
pilhuhn has quit [Quit: Leaving...]
pilhuhn has joined #jruby
<GitHub133>
[jruby] nirvdrum pushed 1 new commit to truffle-head: https://git.io/vXEdp
<GitHub133>
jruby/truffle-head 4d4a170 Kevin Menard: [Truffle] Break dependency on RubyEncoding in StringOperations.
<GitHub25>
[jruby] eregon pushed 1 new commit to truffle-head: https://git.io/vXEF6
<GitHub25>
jruby/truffle-head fd5b2f1 Benoit Daloze: [Truffle] Implement alias of global variables.
<lopex>
nirvdrum: hi there
enebo has joined #jruby
shellac has joined #jruby
<nirvdrum>
lopex: Howdy. I was wondering if you know why there's both ByteList.create and RubyString.createBytelist.
<nirvdrum>
They seem to have slightly different semantics.
<nirvdrum>
Sorry, RubyString#encodeBytelist
<GitHub113>
[jruby] eregon pushed 1 new commit to truffle-head: https://git.io/vXENi
<GitHub113>
jruby/truffle-head b11a300 Benoit Daloze: [Truffle] Replace most calls to createRope by encodeRope.
<lopex>
nirvdrum: encodeBytelist seems to come from headius fix 7ceb0f41140063e36f61785fbdbff0343d64f9de
<lopex>
nirvdrum: if there's two, it's by accident I guess
<lopex>
the one in Bytelist should be much older
<lopex>
wow, the one from bytelist is from 2008 by Ola
<lopex>
nirvdrum: probably when 1.8 was in the wild
<lopex>
I wonder if it should be used at all now
<nirvdrum>
Interesting.
<nirvdrum>
lopex: The other one I was wondering about, but I don't think you wrote, is RubyEncoding.UTF8Encoder.
<nirvdrum>
I'm trying to find out what value it provides over the built-in decoder.
vtunka has quit [Quit: Leaving]
<enebo>
nirvdrum: so I see you committed jnr-ffi
<lopex>
you mean UTF8Coder ?
<enebo>
nirvdrum: fix
<enebo>
nirvdrum: is that good for a release then?
<nirvdrum>
lopex: Yeah. Sorry.
<nirvdrum>
enebo: I still need to test it out. I'll get some time on a server in a little bit.
<enebo>
nirvdrum: ok
<nirvdrum>
It's problematic with older glibc, like you'd find on RHEL or OEL.
vtunka has joined #jruby
<enebo>
someone should deprecate ola create method I think :)
<nirvdrum>
So, unfortunately I can't just test on my machine.
<lopex>
git show 1dcf31ae0a
<lopex>
er
<enebo>
not because ola made it but because it is probably going to make something 1.8-like
<lopex>
digging through blames
<nirvdrum>
enebo, lopex: I've tried to remove it, but symbols break. They must rely on the ISO-8859-1 decoder somehow.
<enebo>
nirvdrum: that is possible if it does 8859-2
<lopex>
nirvdrum: some headius caching commit
<enebo>
err -1
temporalfox has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<enebo>
nirvdrum: well I really really really want to get our symbols fixed for next point release
<enebo>
nirvdrum: I have to get it in right away though after a release which is why I keep slipping
<nirvdrum>
I just can't work out what the semantics should be. ISO-8859-1 and ASCII-8BIT should use the same byte representation, but different interpretations, I believe. So is the idea here to convert Latin-1 characters to UTF-16 codepoints?
<GitHub154>
[jruby] enebo closed pull request #4215: do not eagerly set eval-type in a thread-local (master...test-null-eval-type) https://git.io/vPRq7
<nirvdrum>
enebo: I think that jnr-ffi change is fine. I haven't run the entire test suite, but the likelihood of regression seems incredibly small.
<enebo>
nirvdrum: ok great. Thanks for find and fix
vtunka_ has joined #jruby
temporalfox has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
vtunka_ has quit [Client Quit]
vtunka_ has joined #jruby
vtunka has quit [Ping timeout: 260 seconds]
<nirvdrum>
enebo: I guess the only thing is the released jnr-posix would reference jnr-ffi 2.1.0, while jruby would reference 2.1.1. It should be fine, but something to be aware of.
<enebo>
nirvdrum: yeah I did wonder if we should rev jnr-posix but then we would pull in some other commits
<enebo>
nirvdrum: I don’t want to make a branch for it :)
<nirvdrum>
Not my chair, not my problem. That's what I say.
shellac has joined #jruby
vtunka_ has quit [Quit: Leaving]
vtunka has joined #jruby
donV has quit [Ping timeout: 268 seconds]
<headius>
release day!
<headius>
what do we need releases of, folks?
pawnbox has joined #jruby
<enebo>
headius: jnr-ffi for sure but I tested bits against what is specified in core/pom.rb as stands and saw no issues. So it would only be outstanding issues driving any changes at this point
<headius>
since I think there's updates all across jnr-* I'll start at the bottom and release anything that's changed
<enebo>
headius: but I do not want to pull in anything which might break stuff line unixsocket
<enebo>
headius: so I don’t mind releasing but we should be careful about what we actually put in core.rb
<enebo>
s/line/like/
<headius>
yeah I'll review anything that's not released
<headius>
like I just merged a build fix to jnr-ffi
<enebo>
ok I would jnr-unixsocket is too scary
<enebo>
not because it is broken but just because stuff changed and we don’t get awesome coverage
<nirvdrum>
Heh.
<nirvdrum>
There are some interesting PRs for jnr-posix, but I'd defer those until the next release.
<headius>
yeah
<headius>
I'll review but not merge anything deep
<enebo>
nirvdrum: if they are only additions I am ok but if they change any methods we use then yeah I would prefer to avoid them as well
<headius>
ok, nothing but the build change so I won't spin jffi
<headius>
enebo, nirvdrum: how about the new solaris constants?
<headius>
I think nirvdrum determined we were using all default constants because platform changed from sunos to solaris to match MRI
<headius>
so this *should* be only good
<enebo>
headius: probably. I should not say this but it is not a top-tier OS for us :)
<headius>
I don't know if they got into a snapshot for testing yet but they only add solaris constants alongside sunos ones
<headius>
I'll spin it
<enebo>
headius: but people were getting bogus constants
<headius>
yeah
<enebo>
headius: so likely things probably won’t be any more bad just perhaps not totally fixed
<headius>
other changes to jnr-constants: nl_langinfo constants and missing FreeBSD values from nirvdrum
<headius>
again, not a top-tier platform
<enebo>
yeah I guess worst case if we totally booch those we can re-release next week
<enebo>
but the common platforms won’t be seeing issues
<nirvdrum>
The nl_langinfo constants are additive, but shouldn't really affect anyone.
<nirvdrum>
I suppose the only open question there is if the LangInfo name is fine. I took the name from the header file rather than the function.
<headius>
nirvdrum: seems appropriate to me
<nirvdrum>
The Solaris constants would actually alter behavior for people on that platform. I haven't run the full test suite to tell what the impact is there.
<headius>
what about joni and jcodings?
<nirvdrum>
But, it should fix things.
<headius>
nirvdrum: yeah I think it's more good than risky
<nirvdrum>
The jnr-posix PR for execv looks interesting, but might be risky to shove in last minute.
<nirvdrum>
Otherwise, the only change there is my addition for nl_langinfo.
<GitHub19>
[jruby] eregon pushed 1 new commit to truffle-head: https://git.io/vXuZ8
<GitHub19>
jruby/truffle-head 4217ec8 Benoit Daloze: [Truffle] Update to Truffle 0.19.
<nirvdrum>
Which I still think you might want to consider for Encoding.locale_charmap. But, what you have currently works and I didn't want to add a new native dependency there without discussing with you guys.
<headius>
library search order thing is the bit we want from jnr-ffi yeah?
<nirvdrum>
Yes.
<nirvdrum>
Which really is just a fix for a regression in 2.1.0.
<headius>
only other change is a build fix...ok
<enebo>
execv should not go in :)
temporalfox has joined #jruby
<headius>
indeed
m4rCsi has quit [Quit: No Ping reply in 180 seconds.]
<headius>
nor this struct change in jnr-ffi
<headius>
I will roll those into snapshots after all jnr-* is released and we can let them bake
<GitHub174>
[jruby] headius closed issue #4229: cloned singleton method exists but can't be overwritten https://git.io/vPKBp
<GitHub142>
[jruby] headius closed issue #4228: remove_method fails to remove singleton method on cloned object https://git.io/vPK4E
pawnbox_ has joined #jruby
pawnbox has quit [Read error: Connection reset by peer]
<headius>
enebo: jruby-readline is a nop...I already released it and bumped master in Sept
<headius>
so nothing needed there
<GitHub122>
[jruby] headius closed issue #4141: IRB: arrow keys to move along cmd line and up/down in history not working on Windows https://git.io/vicII
pawnbox_ has quit [Remote host closed the connection]
pawnbox has joined #jruby
pawnbox has quit [Remote host closed the connection]
pawnbox has joined #jruby
bbrowning is now known as bbrowning_away
<GitHub54>
[jruby] nirvdrum pushed 1 new commit to truffle-head: https://git.io/vXug3
<GitHub54>
jruby/truffle-head 09b0660 Kevin Menard: [Truffle] Bump to released versions of jnr-posix and jnr-constants.
<GitHub143>
jruby/master 9187c5e Charles Oliver Nutter: Incorporate MRI's fix for #3886.
<GitHub92>
[jruby] headius closed pull request #3886: Collect all DNS sections by calling to each_resource instead of each_answer (master...resolv-ns-authoritative) https://git.io/vrmEz
<GitHub193>
jruby/master cb5a7be Charles Oliver Nutter: Missed file from RubyGems 2.6.8 update.
<headius>
odd that file wasn't needed to pass tests
drbobbeaty has quit [Ping timeout: 240 seconds]
pil-afk has quit [Quit: Leaving...]
gettalong has joined #jruby
bbrowning_away is now known as bbrowning
pawnbox_ has joined #jruby
<gettalong>
Hi! I developed a library for working with PDF files, HexaPDF (see https://github.com/gettalong/hexapdf). I have been asked how it performs on JRuby but when trying to run the test suite with JRuby 9.1.2.0 it just hangs, consuming no CPU. I tried using "--debug -d -w" to see where it is hanging but to no avail. Any pointers on how to proceed?
kares has quit [Read error: Connection reset by peer]
pawnbox has quit [Ping timeout: 258 seconds]
<chrisseaton>
If you have the time now, can you run jstack (without those flags you mentioned) to tell us where it is hung
kares has joined #jruby
<chrisseaton>
I love the look of hexapdf - it's on my todo list to try it!
<headius>
gettalong: on your test suite I get one error and then it hangs for me a bit later
<headius>
huh, no hang that time
<headius>
3F3E
<headius>
looks like a couple encoding issues, something using ObjectSpace,
<headius>
that might be everything
<gettalong>
headius: Yeah, I tried narrowing it down and it seemed to hang in test/hexapdf/image_loader/test_png.rb but only with certain --seed values
<headius>
ah, that's interesting
<headius>
gettalong: if you find a specific seed that always does it, open an issue and we can try to fix it
<gettalong>
chrisseaton: The main thread hangs at "java.lang.Thread.State: WAITING (parking)" and the backtrace shows "org.jruby.ext.fiber.FiberQueue$1.run(FiberQueue.java:54)" at the last line in JRuby.
<headius>
interesting...Fibers have been pretty stable lately but you might have found another problem
<chrisseaton>
gettalong: are you using fibres directly yourself?
<chrisseaton>
gettalong: ah a kind of coroutine approach
thedarkone2 has joined #jruby
<chrisseaton>
gettalong: fibres aren't as light on JRuby as they are on MRI, but if you are using them fairly coarsely it shouldn't matter
<gettalong>
headius: The command "ruby -Ilib:test test/hexapdf/image_loader/test_png.rb --seed 7918 -v -n '/17/'" seems to hang every time, but if I add the option "--dev" it runs through
<headius>
also interesting
<headius>
could mean something's wrong with the JIT
<headius>
and yes...you will have mixed results with Fibers on JRuby since we have to implement them with native threads
<gettalong>
chrisseaton: Fiber objects are only created on demand from procs. This is done because each Fiber object on MRI uses about 1K of memory and allocating them beforehand lead to massive memory use for PDFs with many stream objects
<gettalong>
headius: Should I open an issue?
<headius>
yes please
<headius>
assuming it still hangs on 9.1.5.0
Specialist has quit [Ping timeout: 260 seconds]
<headius>
hmm, that doesn't hang for me here
<headius>
I'll try it a few times though
<headius>
that might mean it's fixed!
<headius>
:-D
<headius>
I tried it a dozen times and that seed does not hang for me
drbobbeaty has quit [Ping timeout: 265 seconds]
claudiuinberlin has joined #jruby
<GitHub31>
[jruby] francocorreasosa opened issue #4270: Null pointer exception in Rails (devise) https://git.io/vXuPu
<gettalong>
Hmm... I just started writing the issue and running the test suite and *thought* that I found a seed value that always hanged but no, just sometimes again
<GitHub111>
[jruby] gettalong opened issue #4271: JRuby hangs when not using "--dev" mode https://git.io/vXuXG
<headius>
gettalong: freezing with 9.1.5.0?
<gettalong>
Yes, with 9.1.5.0. I did a "rvm get head" followed by a "rvm install jruby" which gave me the JRuby I'm testing with now
<headius>
nirvdrum: oh, you didn't set it to use the snapshot?
<headius>
won't affect JRuby but I'll fix the pom
<nirvdrum>
Unfortunately, not. I updated in jruby since enebo said he wasn't looking to release jnr-posix.
<nirvdrum>
I didn't think of it earlier.
<headius>
ok, well it's in now
tcrawley is now known as tcrawley-away
pilhuhn is now known as pil-afk
pil-afk is now known as pil-zzZZzz
pawnbox has joined #jruby
pawnbox has quit [Ping timeout: 250 seconds]
bbrowning is now known as bbrowning_away
claudiuinberlin has quit []
<GitHub156>
[jruby] nirvdrum pushed 2 new commits to truffle-head: https://git.io/vXzGT
<GitHub156>
jruby/truffle-head e3a43ed Kevin Menard: [Truffle] Memory stats aren't available on AOT.
<GitHub156>
jruby/truffle-head 78ac114 Kevin Menard: [Truffle] Disable call graph tools on AOT since they rely on runtime reflection, which isn't available.
doughathaway has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]