<nirvdrum>
And the version output on that is jruby 9.0.1.0-SNAPSHOT (2.2.2) 2015-08-06 ce91559 Java HotSpot(TM) 64-Bit Server VM 25.60-b23 on 1.8.0_60-b27 +jit [linux-amd64]
<nirvdrum>
There's no 9.0.2.0-SNAPSHOT, but that might just be because it's a nightly.
<headius>
yeah, it would run tonight
<headius>
ahh only the dist build is old
<headius>
snapshots/master/jruby-bin-9.0.1.0-SNAPSHOT.tar.gz for example is from yesterday
<headius>
I don't remember what's in the dist version of that...I think it's an unrenamed tarball or something
<headius>
bin, src, gem, and complete jar are all updated yesterday
<headius>
yeah I think this was just when I was experimenting to get the nightly build to build everything
<headius>
enebo runs a script that renames some things and it probably didn't run for one of the builds when I was fiddling with it
<headius>
I'll remove those from snapshots
<nirvdrum>
So which one should I be using? I just want whatever most closely resembles the distribution tarballs.
<headius>
bin
<headius>
that's what our download links go to for releases
<nirvdrum>
Okay, thanks.
<headius>
sorry for the confusion :-)
<nirvdrum>
No worries. An old cron job that I just realized was running a really out of date build.
erikhatcher has quit [Quit: erikhatcher]
temporalfox has quit [Read error: Connection reset by peer]
temporalfox has joined #jruby
erikhatcher has joined #jruby
<headius>
ok
erikhatcher has quit [Client Quit]
camlow325 has quit []
<GitHub44>
[jruby] brauliobo opened issue #3305: [Truffle] Error on build http://git.io/vGDQJ
dinfuehr has joined #jruby
havenwood has joined #jruby
dinfuehr has quit [Ping timeout: 250 seconds]
benlovell has joined #jruby
benlovell has quit [Ping timeout: 240 seconds]
pitr-ch has quit [Read error: No route to host]
dbussink has quit [Ping timeout: 256 seconds]
dbussink has joined #jruby
dinfuehr has joined #jruby
dinfuehr has quit [Ping timeout: 255 seconds]
mdedetrich has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
mdedetrich has joined #jruby
benlovell has joined #jruby
benlovell has quit [Ping timeout: 244 seconds]
dinfuehr has joined #jruby
subbu has joined #jruby
dinfuehr has quit [Ping timeout: 264 seconds]
decentrality has quit [Read error: Connection reset by peer]
decentrality has joined #jruby
dinfuehr has joined #jruby
qmx has quit [Ping timeout: 246 seconds]
qmx has joined #jruby
qmx has quit [Changing host]
qmx has joined #jruby
_djbkd has joined #jruby
skade has joined #jruby
havenwood has quit [Remote host closed the connection]
dinfuehr has quit [Remote host closed the connection]
skade has quit [Quit: Computer has gone to sleep.]
skade has joined #jruby
skade has quit [Client Quit]
_djbkd has quit [Quit: My people need me...]
subbu has quit [Quit: Leaving]
yfeldblum has quit [Ping timeout: 244 seconds]
dinfuehr has joined #jruby
donV has quit [Quit: donV]
skade has joined #jruby
dinfuehr has quit [Remote host closed the connection]
pietr0 has quit [Quit: pietr0]
dinfuehr has joined #jruby
donV has joined #jruby
nirvdrum has quit [Ping timeout: 272 seconds]
skade has quit [Quit: Computer has gone to sleep.]
yfeldblum has joined #jruby
yfeldblu_ has joined #jruby
yfeldblum has quit [Ping timeout: 244 seconds]
yfeldblu_ has quit [Ping timeout: 252 seconds]
donV has quit [Quit: donV]
rsim has joined #jruby
yfeldblum has joined #jruby
yfeldblum has quit [Remote host closed the connection]
yfeldblum has joined #jruby
yfeldblu_ has joined #jruby
yfeldblum has quit [Ping timeout: 244 seconds]
temporal_ has joined #jruby
temporalfox has quit [Ping timeout: 272 seconds]
mkristian_ has joined #jruby
mkristian__ has joined #jruby
mkristian_ has quit [Ping timeout: 250 seconds]
yfeldblu_ has quit [Remote host closed the connection]
yfeldblum has joined #jruby
benlovell has joined #jruby
samphippen has joined #jruby
benlovell has quit [Ping timeout: 255 seconds]
temporalfox has joined #jruby
temporal_ has quit [Ping timeout: 246 seconds]
robbyoconnor has joined #jruby
shellac_ has joined #jruby
shellac_ has quit [Quit: Computer has gone to sleep.]
shellac_ has joined #jruby
temporalfox has quit [Read error: Connection reset by peer]
temporalfox has joined #jruby
skade has joined #jruby
shellac_ has quit [Quit: Computer has gone to sleep.]
skade has quit [Ping timeout: 246 seconds]
mkristian__ has quit [Quit: This computer has gone to sleep]
samphippen has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
_djbkd has joined #jruby
dinfuehr has quit [Remote host closed the connection]
_djbkd has quit [Quit: My people need me...]
benlovell has joined #jruby
mdedetrich has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
benlovell has quit [Ping timeout: 250 seconds]
skade has joined #jruby
benlovell has joined #jruby
shellac_ has joined #jruby
pitr-ch has joined #jruby
samphippen has joined #jruby
<GitHub143>
[jruby] eregon commented on commit afa0919: Would `flat_map` be a good fit here? http://git.io/vGSIJ
<ebarrett>
chrisseaton: hey
<chrisseaton>
ebarrett: hello
<ebarrett>
sorry about the dealy with the graal bug
<ebarrett>
took a while to figure out how to repro
<chrisseaton>
np - I'll try today
<ebarrett>
thanks
dinfuehr has joined #jruby
vtunka has joined #jruby
dinfuehr has quit [Ping timeout: 264 seconds]
shellac_ has quit [Quit: Computer has gone to sleep.]
<GitHub125>
[jruby] pitr-ch commented on commit afa0919: :+1: http://git.io/vGS37
digitalextremist has quit [Remote host closed the connection]
digitalextremist has joined #jruby
shellac_ has joined #jruby
shellac has quit [Quit: Ex-Chat]
shellac_ is now known as shellac
<GitHub70>
[jruby] chrisseaton commented on commit 16b8aed: @pitr-ch ping http://git.io/vGS8Q
samphippen has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
samphippen has joined #jruby
dinfuehr has joined #jruby
mkristian__ has joined #jruby
mkristian_ has joined #jruby
mkristian__ has quit [Ping timeout: 246 seconds]
dinfuehr has quit [Ping timeout: 256 seconds]
skade has quit [Ping timeout: 264 seconds]
shellac has quit [Quit: Computer has gone to sleep.]
cristianrasch has joined #jruby
shellac has joined #jruby
<GitHub73>
[jruby] chrisseaton pushed 1 new commit to truffle-java-interop: http://git.io/vGSVI
<GitHub73>
jruby/truffle-java-interop 44be094 Chris Seaton: [Truffle] Add a .static_call method on foreign objects.
mdedetrich has joined #jruby
samphippen has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<GitHub160>
[jruby] chrisseaton pushed 1 new commit to truffle-head: http://git.io/vGSr0
<GitHub160>
jruby/truffle-head 59bfc0b Chris Seaton: Merge branch 'master' into truffle-head
shellac has quit [Quit: Computer has gone to sleep.]
mdedetrich has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
shellac has joined #jruby
<GitHub43>
[jruby] pitr-ch pushed 4 new commits to master: http://git.io/vGS6v
<GitHub43>
jruby/master e8610d7 Petr Chalupa: [Truffle] jruby+truffle_runner: change default bundle path to .jruby+truffle_bundle
<GitHub43>
jruby/master 69231c8 Petr Chalupa: [Truffle] jruby+truffle_runner: do not polute stdout
<GitHub43>
jruby/master b2d8260 Petr Chalupa: [Truffle] jruby+truffle_runner: use flat_map
<GitHub105>
[jruby] eregon pushed 2 new commits to master: http://git.io/vGSXf
<GitHub105>
jruby/master e4218fa Benoit Daloze: [Truffle] Actually raise the error.
<GitHub105>
jruby/master 67900d0 Benoit Daloze: [Truffle] Removed the unused rubyWithSelf().
cristianrasch has quit [Ping timeout: 250 seconds]
dinfuehr has joined #jruby
drbobbeaty has joined #jruby
dinfuehr has quit [Ping timeout: 268 seconds]
mkristian_ has quit [Quit: This computer has gone to sleep]
pitr-ch has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
mkristian_ has joined #jruby
mkristian__ has joined #jruby
mkristian_ has quit [Ping timeout: 250 seconds]
skade has joined #jruby
yfeldblum has quit [Ping timeout: 246 seconds]
shellac has quit [Quit: Computer has gone to sleep.]
nirvdrum has joined #jruby
cristianrasch has joined #jruby
shellac has joined #jruby
<GitHub46>
[jruby] nirvdrum commented on commit 7a99f1c: Small typo: "caling" -> "calling" http://git.io/vGSx3
dinfuehr has joined #jruby
<GitHub89>
[jruby] chrisseaton commented on commit 7a99f1c: It would be good if PRs let you edit the file being reviewed inline, which would create a new commit on the PR branch http://git.io/vGSpr
bbrowning_away is now known as bbrowning
dinfuehr has quit [Ping timeout: 256 seconds]
havenwood has joined #jruby
pitr-ch has joined #jruby
pitr-ch has quit [Client Quit]
skade has quit [Quit: Computer has gone to sleep.]
erikhatcher has joined #jruby
mdedetrich has joined #jruby
pitr-ch has joined #jruby
mdedetrich has quit [Read error: Connection reset by peer]
pitr-ch has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
pitr-ch has joined #jruby
pitr-ch has quit [Client Quit]
lance|afk is now known as lanceball
pitr-ch has joined #jruby
dinfuehr has joined #jruby
tcrawley-away is now known as tcrawley
dinfuehr has quit [Ping timeout: 252 seconds]
arturaz has joined #jruby
<rtyler>
o/
Aethenelle has joined #jruby
pitr-ch has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
pitr-ch has joined #jruby
pitr-ch has quit [Client Quit]
vyorkin has joined #jruby
pitr-ch has joined #jruby
skade has joined #jruby
mkristian__ has quit [Quit: This computer has gone to sleep]
pitr-ch has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
enebo has joined #jruby
zapho53 has joined #jruby
vyorkin has quit [Quit: WeeChat 1.2]
pitr-ch has joined #jruby
vyorkin has joined #jruby
<zapho53>
Thinking of investing a lot of time developing an app in JRuby and wondered if anyone can give me some rough estimates of JRuby 9000's potential for performance optimisation in the next 2 years of its development. The app will run inside Torquebox so how would that compare with MRI/Unicorn?
benlovell has quit [Ping timeout: 255 seconds]
pitr-ch has quit [Client Quit]
vyorkin has quit [Read error: Connection reset by peer]
dinfuehr has joined #jruby
<nirvdrum>
headius: I'm not seeing any 9.0.2.0-SNAPSHOTs in http://ci.jruby.org/. Did we just miss the window yesterday?
<headius>
good morning!
<headius>
hmm
<headius>
well it was red in the most recent build
<headius>
I'll check on that today and beat it into shape
<headius>
it's that blasted EBADF
<headius>
I'll try another build, but we might want to get master building with a 9k release now
dinfuehr has quit [Ping timeout: 264 seconds]
<nirvdrum>
Heh, sorry. I should have lead with a greeting :-)
<nirvdrum>
And I didn't realize our nightlies were tied to build status. Good to know.
skade has quit [Quit: Computer has gone to sleep.]
mkristian__ has joined #jruby
vyorkin has joined #jruby
<headius>
well, the dist has to complete at least
mkristian_ has joined #jruby
<headius>
that's what failed
<headius>
mkristian_: just pinged you in gitter
<headius>
dunno if you're really here or not
<headius>
:-)
camlow325 has joined #jruby
mkristian__ has quit [Ping timeout: 240 seconds]
<mkristian_>
yes I am pretty happy that all those EBADF are gone with jruby-9k. we have to configure each gem-maven-plugin with :jrubyVersion => '9.0.1.0'.
<headius>
zapho53: Torquebox compares very favorably with those servers, and 9k has great potential for perf improvements over just the next six months
<headius>
already folks have seen improvements over 1.7.x, which was generally faster than CRuby already
<headius>
mkristian_: ok
<rtyler>
headius: are any of us really here?
<mkristian_>
headius, but it can be a lot of place. a jruby.version property in ./pom.rb could do the trick - not sure if this has issues with tests.
<zapho53>
headius: I haven't used Torquebox yet but does its threading model result in significantly reduced memory usage with large apps compared with Unicorn/MRI?
<headius>
rtyler: deep thoughts
<headius>
zapho53: it can run your applications with a single JRuby instance handling many threads at once, yes
<headius>
the base memory use for a single threaded app will be higher, but it pays off after 2 or 3 instances
<nirvdrum>
headius, mkristian_: Unrelated to the issue at hand, would it be possible to build the jruby-truffle JAR with JDK 8 and still assemble a dist that uses JDK 7 for the rest?
<headius>
or rather it pays off compared to 3+ instances of unicorn/MRI
erikhatcher has quit [Quit: erikhatcher]
<headius>
mkristian_: I will try that and see what happens
<zapho53>
headius: Do you mean 1 TB instance would use about the same memory as 3 Unicorns?
<headius>
something like that
<bbrowning>
zapho53: it all hinges greatly on the memory usage patterns of your app
<mkristian_>
nirvdrum, you probably just need to set target and source version for the compiler
<bbrowning>
but, in general, as the # of MRI processes go up the total memory usage will surpass that of something on jruby using threads
<nirvdrum>
mkristian_: And that won't mess things up if enebo builds the next dist starting with JDK 7?
<bbrowning>
for most (every?) app there's a tipping point where jruby threading model will use less memory than MRI process model
<nirvdrum>
mkristian_: I also wasn't sure what that does for any uber JARs we have.
<headius>
I'd assume for every app
camlow32_ has joined #jruby
<headius>
nirvdrum: ci dist worked this time
pitr-ch has joined #jruby
<mkristian_>
nirvdrum, the truffle jar is still separated from any other build artifact, i.e. gets included as jar.yes the downside all build needs to use jdk8 !
<ebarrett>
chrisseaton: any joy?
<mkristian_>
truffle is just part of the main build
<headius>
nirvdrum: what's the problem?
<chrisseaton>
ebarrett: sorry haven't had a chance yet - going to give it a go at some point today but can't guarantee
<nirvdrum>
mkristian_: Do you think that's a problem in practice?
<ebarrett>
chrisseaton: no worries
pitr-ch has quit [Client Quit]
camlow325 has quit [Ping timeout: 244 seconds]
<headius>
enebo gets tweaky about building with a newer JDK because something always creeps in with newer bytecode
<headius>
we can fix those but sometimes we don't see them until they're in the wild
<zapho53>
headius: Any estimates on when you reckon Truffle/Graal might be ready for production apps? I know it's all a bit speculative but, hey ..
<nirvdrum>
headius: No problem per se. It just seemed that since the Truffle activation is completely opt-in, we might be able to get away with going 8+.
<mkristian_>
nirvdrum, well running travis with java-7 will be an issue
<nirvdrum>
If that's undesirable, I won't pick at that thread.
<enebo>
heh
<chrisseaton>
zapho53: not an easy question to answer, maybe a year? I think we should be running substantial apps by next RubyConf
<headius>
enebo: I get tweaky too but perhaps not as much :-)
<nirvdrum>
But if it were doable without too much consternation, it'd give the Truffle part a bit more flexibility.
<zapho53>
headius: What is it about 9k which creates the opportunity for new performance optimisations?
<headius>
zapho53: the new compiler allows us to do optimizations above the JVM level to generate better code for JVM
Aethenelle has quit [Quit: Aethenelle]
<zapho53>
chrisseaton: That soon? Encouraging.
camlow32_ has quit [Remote host closed the connection]
<mkristian_>
nirvdrum, let me see if I can "pick" some java8-compiler for truffle only
<enebo>
as long as we can build locally with Java 7 (sans truffle) it is ok. Once we say we have to always build with 8 we will start sneaking in 8 APIs
<headius>
lik e we should be able to do our own inlining, type specialization, etc
<mkristian_>
not sure if this is possible
<nirvdrum>
headius, enebo: More selfishly, JRuby is the only Truffle language that requires Java 7 and people break it with some regularity (both the Truffle team and us), by inadvertently using JDK 8 classes or methods.
<mkristian_>
well, I can add the truffle module via a jdk7 profile. that is easy
<headius>
mkristian_: install_gems still used 1.7.20 :-\
<enebo>
nirvdrum: So you guys want java 8 for truffle which is fine but I am not fine with us support Java 7+ and only being able to use Java 8 to build it for reasons headius stated.
<enebo>
nirvdrum: so full build of all modules can use Java 8 so truffle can compile but when I build -pl core or non-truffle I want it to work with Java 7
<nirvdrum>
enebo: Right. That's essentially how we keep breaking things :-/
<enebo>
nirvdrum: not sure why I am replying to you :)
<mkristian_>
headius, ok that one is coming from the maven polyglot extension
zapho53 has quit [Quit: BitchX: may cause birth defects]
<nirvdrum>
enebo: So what I was wondering is if we could start with JDK 7 and have maven switch to JDK 8 just for that profile/module/whatever.
<enebo>
nirvdrum: yeah. if truffle profile can just switch to 8 that is cool?
<headius>
mkristian_: is there a way to tell it to use newer jruby?
<nirvdrum>
enebo: So even if there is a class file version change, it'd be lazily loaded anyway. And we could do a JVM version check in the launcher to give a sensible error message if using -X+T on Java 7.
<enebo>
nirvdrum: not sure how maven nor magnets work
<nirvdrum>
enebo: That's why we need mkristian_!
<headius>
enebo: that would require everyone building JRuby master to have both 7 and 8
<headius>
e.g. rvm
<nirvdrum>
headius: Ahh, crap. I forgot RVM builds from source.
<nirvdrum>
I don't quite get why.
<headius>
a hard requirement on 8 across the board would also cause issues for any platforms that don't have an 8 release out of the box, like ubuntu 14.04
<headius>
nirvdrum: well, only for jruby-head
<headius>
but people try out jruby-head
<nirvdrum>
Yeah. ruby-build just uses the prebuilt dists.
<headius>
and we often need to get people to do local builds to test patches
<mkristian_>
headius, nirvdrum but with jdk7 it just will not build truffle but everything else
<headius>
asking them to have both 7 and 8 installed is too much, I think
<nirvdrum>
Yeah, you're probably right.
<nirvdrum>
Oh well. It was an interesting though experiment.
<mkristian_>
i.e. it also means truffle will not be avaibable to people running jdk7
<headius>
mkristian_: which wouldn't be useful anyway, really
<headius>
hmm
<headius>
I dunno, it still feels sketchy
<mkristian_>
headius, will look up if there is a way to tell polyglot maven to use another jruby.
<nirvdrum>
Huh. I didn't realize Ubuntu 14.04 didn't ship with JDK 8. I guess I just fixed that problem a while back.
<headius>
mkristian_: ok, I don't see anything obvious in the mvn extension config
<enebo>
headius: yeah I guess so. I just want to be able to build non-truffle with 7. If truffle requires 8 then that is ok with me I just don’t want all paths through 8 if we support 7 (which you know)
<headius>
right
<enebo>
It is good that we talk about ideas before implementing them :)
<nirvdrum>
I guess one way to deal with it is if maven could be instructed to build Truffle only if there's a JDK 8 on the system as well. But that may be getting too fancy.
<nirvdrum>
That would keep RVM users happy at the expense of potential confusion. I think the launcher could detect that case and help out though.
<headius>
it might be worth a try... jdk7 build would be faster then
<nirvdrum>
But this is all starting to sound a lot harder than what I was thinking initially.
<headius>
yeah
<headius>
too many levels of complication for my taste right now
<mkristian_>
headius, it is not there. there is a way to use a given jruby_home but not switch jruby versions
<headius>
it might be possible to get truffle to build using the eclipse java 8 compiler
<headius>
that would be simpler
<headius>
mkristian_: bummer :-\ how risky would it be to make it use 9.0.1.0 now?
<headius>
I guess building 1.7.x on Java 6 would be out but I don't think enebo does that anymore anyway
<headius>
enebo: ?
<enebo>
headius: maven explodes with 6
<nirvdrum>
This may feed into a larger discussion about when JRuby as a whole goes Java 8+. It's been talked around for a while and now seems to be at an indeterminate 10k release. It'd probably be good to get a better roadmap for that.
<enebo>
headius: something goes wrong uploading to sonatype I think
<enebo>
headius: which bugs me but it is what it is
<headius>
nirvdrum: oracle's schedule is rather ambitious given what's supported on most OSes right now
<nirvdrum>
enebo: Maybe SNI? Java 6 didn't support that.
<mkristian_>
headius, I still have problems using embedded jruby-9k with OSGi containers. I rather at the feature to switch jruby version via a property.
<headius>
we know multiple users that have no plans to jump to 8 yet
<nirvdrum>
headius: Above my pay grade :-)
<enebo>
nirvdrum: it was flaky whatever was wrong
<headius>
mkristian_: ahh, I see
<headius>
I'm hoping we'll get a better feel for whether people are on 7 or 8 at JavaOne
<headius>
I suspect there's a lot of people that are still on 7
<mkristian_>
headius, for example eclipse has maven-polyglot support and it uses OSGi.
<headius>
mkristian_: do you know what the issue is?
<nirvdrum>
headius: I wonder if we could get that info from New Relic. I think they do an annual report on "state of the stack".
<headius>
yeah true
<headius>
they did one fairly recently, let me find it
<nirvdrum>
It's obviously biased, but at least it's numbers.
<headius>
I don't consider new relic customers to be indicative of big JVM shops though
<enebo>
I suspect there are massive numbers of Java 7 users
<headius>
I might be wrong about that
<nirvdrum>
I suspect a lot of desktop users have just upgraded due to security issues & Java 7 being EOL'd.
<enebo>
nirvdrum: JRE vs JDK here too
<enebo>
and server vs desktop
<nirvdrum>
Maybe codefinger has numbers he can share on the heroku side?
<nirvdrum>
enebo: Indeed.
<enebo>
I bet money most java devs have jre8 installed but still use jdk 7 for linux jre 7 deploy
vyorkin has quit [Read error: Connection reset by peer]
<headius>
48.5% 7, 38% 8, 13.5% 6 in this poll around May
<enebo>
headius: not very surprising to me
<headius>
yeah, that's about what I'd expect
<enebo>
headius: I figured 7 would win and there would be some Java 6 still :)
<headius>
it will slowly slide toward 8 over the next year
arturaz has quit [Remote host closed the connection]
<enebo>
headius: actual numbers I don’t think I had any clue
<enebo>
headius: yeah 8 will pass 7 in a year but it will probably just flop
<headius>
surprisingly few recent results for "java 8 adoption"
<enebo>
50% 8 and 35% 7
<enebo>
There is a long tail on this stuff
<headius>
"Once-faltering Java is beginning to run away with the Tiobe language popularity index, with the language getting a shot in the arm from last year's Java 8 release."
<headius>
Krill article from Aug 7
<enebo>
COCAINE IN THE ARM
<headius>
that's worth remembering... 8 has only been out for 1.5 years eh?
<headius>
a long time in bleeding edge for a blink of an eye in most orgs wrt big tech decisions
<headius>
for=but
<enebo>
1.5 years is pretty short if your app is running fine and you are not doing much active development
<headius>
"Last year's Java 8 was the crack cocaine Java needed to get its fix"
<enebo>
Even if you are actively developing you might not be that excited about the new stuff
<nirvdrum>
The next Ubuntu LTS release will be in April. I'd imagine they'll go JDK 8 for that. But 14.04 will be supported until 2019.
donV has joined #jruby
<enebo>
Most people do not like to change their builds
<headius>
nirvdrum: yeah I'd expect to see 8 too
<enebo>
I keep thinking most web apps are low resource and low traffic on the web
benlovell has joined #jruby
<headius>
April seems so far away wrt this discussion :0\
<headius>
:-\
<headius>
sigh
<enebo>
So why update to something which runs something running ok to be faster
<enebo>
just another round of testing amirite :)
<headius>
hah right
<chrisseaton>
I've never done ops so I'm probably ignorant, but what's the bothering in installing JDK 8 via a package manager or even just untarring it in a provision script?
<headius>
chrisseaton: it's not in the supported package lists
<chrisseaton>
I get that the package may not be 'supported' by the distro, but are Canonically really going to fix HotSpot bugs?
<nirvdrum>
I guess part of it, too, is I don't know what % of users embed JRuby or use the scripting container stuff. I think those purely using JVM for JRuby are more open to upgrading. And indeed, the promise of indy may have pushed more up to Java 8 already.
<headius>
chrisseaton: at least in the case of Red Hat, OpenJDK (in the form of IcedTea) is definitely part of the contract
<headius>
I don't know how Canonical handles that
<headius>
but yeah, you go outside the supported packages you're on your own if something breaks
<nirvdrum>
They use IcedTea builds as well, but no support contract as far as I know.
<nirvdrum>
The benefit is ease of distribution. And if you install any package that depends on Java, you'll want something registered in the package manager.
<enebo>
chrisseaton: you are also thinking about this logically as well
<headius>
hah, yeah
<headius>
this about it with irrational fear
<enebo>
chrisseaton: larger orgs chant “support” and then never actually use it
<headius>
think
<nirvdrum>
E.g., you can't just install ant, maven, or tomcat from the package manager without having a dependent JDK install as well.
<headius>
neck to choke
<chrisseaton>
I need to do a sabbatical where I go to a company and actually build something with Ruby and deploy it
<enebo>
headius: aka someone to sue
<headius>
it's also a matter of security releases...even if you find an unsupported package archive with 8, how often does it update?
<enebo>
headius: which also never happens
<headius>
and after the Java 7 security debocle, people are tweaky
<nirvdrum>
I don't really understand why Oracle rescinded its distribution license. I think that complicated a lot of things.
<headius>
this stuff is annoying...is it beer o'clock yet?
<enebo>
yeah a few day 0’s keep people afraid for quite a while
mkristian_ has quit [Quit: This computer has gone to sleep]
<enebo>
headius: I will make a yeast starter this morning
<enebo>
headius: it is like making beer
<enebo>
headius: but in a bunsen burner
<headius>
enebo: beer o'clock in the long now
<nirvdrum>
enebo: Well, I think that's the argument for using Oracle releases. CPU builds, while open source, aren't tagged forests or something silly.
benlovell has quit [Ping timeout: 240 seconds]
<enebo>
I would manual installs of anything probably almost never get regularly updated so there is something to be said for packaged solns
<nirvdrum>
So you do do something like "sudo add-apt-repository ppa:webupd8team/java" and it adds an apt source and now packages from that repo can be installed via apt-get, with its own release schedule independent of Canonical's security release schedule.
<enebo>
nirvdrum: I mean I still see the value in this idea overall
<nirvdrum>
It's a way to get more recent software on an otherwise stable system.
<headius>
just kinda curious about rough adoption numbers
bbrowning is now known as bbrowning_away
<codefinger>
ah, yea. we have those. I think i can share relative numbers. let me find out.
<codefinger>
i have worked with a ton of people to upgrade from JDK7 to JDK8 and never had an issue
<nirvdrum>
Someone on the webupd8team (they're about as official as you get with PPAs) repackages with the newest CPU and it comes it same day as the Oracle release.
<enebo>
nirvdrum: may Alin Andrei never tire of putting these out :)
<nirvdrum>
Whereas, based on my experience from the JDK list, IcedTea needs to figure out the changeset in the CPU release and create their own forest, separate QA, etc.
<nirvdrum>
So the IcedTea releases tend to lag behind the Oracle releases for a bit.
<enebo>
codefinger: yeah I would guess 7 to 8 is ok except for some tooling (I have had issues running hprof agent with invokedynamic) but I think people are afraid something could break
<enebo>
nirvdrum: I bet
<nirvdrum>
headius: Correct. Although in this case, it essentially just wgets the tarball from Oracle and explodes on disk, moving files into whatever layout Oracle uses for its openjdk packages.
<enebo>
codefinger: part of earlier discussion was fear of upgrading a working app because theoretically it could cause pain vs expect pain
<headius>
6 to 7 was a much harder update than 7 to 8
<codefinger>
enebo: yea, i went through this when I changed the default JDK on heroku. it was pretty painless
benlovell has joined #jruby
<enebo>
headius: you ever see MH explosion trying to cpu=times with Java 8?
<headius>
codefinger: larger question is what impact it would have if we suddenly said jruby 9k now needed 8
<headius>
enebo and I suspect it would be a pretty big problem
dinfuehr has quit [Ping timeout: 240 seconds]
<headius>
MH explosion?
<enebo>
headius: hprof not working
<headius>
yes
<headius>
on some releases I have
<headius>
they fixed a few issues
<enebo>
headius: codefinger: I suspect Heroku has a lot more hip current users who are more likely to be amenable to upgrades versus back office intranet deploys
<nirvdrum>
headius: So, yeah, it's all open source and they don't repackage anything. You can see the dpkg itself is in the KB range.
<codefinger>
enebo: no comment...
<enebo>
hahah
<enebo>
I expected that reply
<enebo>
nirvdrum: if I was using linux on daily basis I would use this instead of manually downloading
<nirvdrum>
headius: It probably still makes sense to do it for JRuby 10k. I'm just suggesting having a rough timeline would make sense. Should it be a year after Java 7 has been EOL'd? Should it be tied to the Java 9 release? Should it be tied to OS vendor releases?
<enebo>
nirvdrum: too bad EA builds are publshed here
<enebo>
are not
<nirvdrum>
headius: And then that snowballs into a discussion about things like Ruboto.
<enebo>
I see beta builds I guess
<nirvdrum>
enebo: They have JDK 9 builds.
<headius>
nirvdrum: oh sure, for 10k I don't think that's a question
<headius>
Android is now the lagging-behind mobile platform ME was before it died
<nirvdrum>
headius: Of course, I'd like to see a narrower window than 3 years for 9k -> 10k :-D
Aethenelle has joined #jruby
<headius>
hah...I'd be surprised
<headius>
9k was a huge move
<enebo>
To some degree what version we supports depends on figuring out what people use and how painful backwards compat is
<enebo>
I think Java 8 closures would make some parts of out code much cleaner but I have never changed any of that code to make sure it still would run well with them
<nirvdrum>
enebo: But anything older than 6 months is legacy to Rubyists!
<enebo>
nirvdrum: yet you see the wasteland of never updated 0.9 rails apps litter the planet
subbu has joined #jruby
<nirvdrum>
Heh.
<enebo>
nirvdrum: you are right of course. Ruby users seem to like change
<nirvdrum>
Well, that's because Rails got slower with subsequent releases. I stuck with 2.3 as long as I could.
<enebo>
nirvdrum: the idea of moving the bar higher is enough reason to change
<enebo>
nirvdrum: I think that was arel which killed you right?
<nirvdrum>
I ultimately moved on because I was switching to JRuby and Rails 2.3's multithreading story was scattershot.
<enebo>
nirvdrum: seeing some of aaron’s profiling since 3.x has given me some confidence they have improved perf past 3.0
<nirvdrum>
But, Arel performance was pretty bad. Worse was ActiveRecord not really having a plugin API, so useful things just didn't work with AR-JDBC. Sequel turned out to be a lot faster, has a sane (read: actual) plugin API, and a regular release schedule.
<enebo>
nirvdrum: we have never heard of anyone moving from Sequel to AR :)
<enebo>
nirvdrum: sounds like memory is the other major win with moving to Sequel
<nirvdrum>
The tangential benefit is since AR really doesn't have an API -- it's all monkey patches -- upgrading Rails became a lot easier when I switched to Sequel.
<nirvdrum>
My own benchmarks had shown that call where() was substantially faster than the find_by_XXX calls because it doesn't need to go through method_missing.
<enebo>
nirvdrum: yeah I was at that railsconf when they landed this
<nirvdrum>
*calling
pjungwir has joined #jruby
mkristian_ has joined #jruby
<pjungwir>
Good morning! I have a Rails app in JRuby 9.0.0.0, and I'm researching how to package it for IBM Websphere. I was wondering how close warbler 2.0 is to being released?
mkristian__ has joined #jruby
<pjungwir>
I might be able to get paid to contribute myself or have someone else paid to work on it,
<pjungwir>
but I'd like to know if it's like 1 month out or 6 months out?
<pjungwir>
Also should I be scared by the 111 open issues?
<pjungwir>
Is there a different approach to running a JRuby Rails app in a J2E-centric organization?
<headius>
pjungwir: codefinger is a good one to talk to about helping with warbler 2
mkristian_ has quit [Ping timeout: 260 seconds]
<pjungwir>
Okay, thank you! Via IRC or some other way?
<headius>
he's here now, but I'll let him speak for how he'd like to communicate :-)
camlow325 has joined #jruby
camlow325 has quit [Remote host closed the connection]
<codefinger>
hello
<codefinger>
pjungwir: warbler 2 is looking pretty god
<pjungwir>
Good morning! Was just looking at the table of contents for your Deploying with JRuby book. :-)
<headius>
better pretty god than ugly god
<codefinger>
pjungwir: oh, that book is out of date and out of print now. :(
<pjungwir>
Oh well, I might read it anyway. :-)
<codefinger>
pjungwir: there's a rumor that there will be a new version sometime next year ;)
<pjungwir>
Anyway, I was wondering if you'd like help on warbler 2, and roughly what kind of effort it would take to make it production-ready for jr9k?
<pjungwir>
Ha ha, okay, I'll watch for it!
<codefinger>
pjungwir: warbler 2 is in pretty good shape. I've been using it for some thing, and I know others have too. I should probably just pull together the first RC or something
<codefinger>
pjungwir: if you could just give it a test, that would be great.
<pjungwir>
Okay.
<pjungwir>
I'm sort of worried about 111 open issues. A lot of them seem spurious, but still. . . .
<codefinger>
pjungwir: i'm not sure what's needed for Websphere, but feel free to hit me with questions
<codefinger>
pjungwir: eh, most of those were inherited from ages ago. not even sure if their are relevant anymore
<codefinger>
s/their/they're/
<pjungwir>
That's true, only a few are from 2015.
<pjungwir>
Okay, so I'll give it a try and let you know what troubles I run into. I'm happy to contribute code if that would be helpful.
<pjungwir>
Btw do you have any advice for handling delayed_job in a J2E context?
<pjungwir>
"Buy the book"? :-)
<pjungwir>
Is there some J2E service I can use instead for ad hoc background tasks?
<pjungwir>
Or maybe multithreading makes it an easier problem to solve.
<pjungwir>
get_back looks like it's probably what I want. Thanks!
<pjungwir>
One more question okay?
<codefinger>
yea, no prob
<codefinger>
i'm tabbing in and out, but i'm here
<pjungwir>
For scheduled periodic jobs (not ad hoc ones) we are using whenever/cron to run rake tasks. I'm not sure how I'd run a rake task from a war file.
<pjungwir>
But I'm sure from my J2EE days there is some way to schedule code to run.
<codefinger>
pjungwir: oh but you can! take a look at the `runnable` feature
<pjungwir>
Okay, that's great!
<codefinger>
pjungwir: you can do something like `java -jar myapp.war -S rake task`
<pjungwir>
Thanks for all your help. I think you've already saved me hours of Googling.
shellac has quit [Quit: Computer has gone to sleep.]
benlovell has joined #jruby
rcvalle has joined #jruby
pitr-ch has joined #jruby
vyorkin has joined #jruby
rsim has quit [Quit: Leaving.]
mje113 has joined #jruby
camlow325 has quit [Remote host closed the connection]
samphippen has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
camlow325 has joined #jruby
<GitHub104>
[jruby] mkristian opened issue #3306: psych can not load yaml file from classloader http://git.io/vGHNt
cajone has joined #jruby
camlow325 has quit [Remote host closed the connection]
rsim has joined #jruby
dinfuehr has joined #jruby
dinfuehr has quit [Ping timeout: 244 seconds]
robbyoconnor has quit [Ping timeout: 268 seconds]
robbyoconnor has joined #jruby
tcrawley is now known as tcrawley-away
camlow325 has joined #jruby
camlow325 has quit [Remote host closed the connection]
samphippen has joined #jruby
benlovell has quit [Ping timeout: 264 seconds]
oblutak has joined #jruby
havenwood has joined #jruby
pitr-ch has quit [Ping timeout: 264 seconds]
_djbkd has joined #jruby
robbyoconnor has quit [Ping timeout: 260 seconds]
mkristian__ has quit [Quit: This computer has gone to sleep]
shellac has joined #jruby
bb010g has joined #jruby
subbu is now known as subbu|lunch
camlow325 has joined #jruby
pitr-ch has joined #jruby
dinfuehr has joined #jruby
dinfuehr has quit [Ping timeout: 256 seconds]
lance|afk is now known as lanceball
vyorkin has quit [Quit: WeeChat 1.2]
yfeldblum has joined #jruby
<GitHub191>
[jruby] mkristian opened issue #3307: jruby-complete-1.7.x comes with yaml ruby files from yecht.jar which conflicts with stdlib http://git.io/vGQKT
yfeldblum has quit [Remote host closed the connection]
subbu|lunch is now known as subbu
<xardion>
Dunno if this plays into the earlier conversation about JDK 7 & 8, but I do run my jruby app on Java 8 without issue
donV has quit [Quit: donV]
dinfuehr has joined #jruby
subbu is now known as subbu|away
<GitHub78>
[jruby] pitr-ch commented on commit 7a99f1c: thanks, coming with next changes http://git.io/vGQ7E
dinfuehr has quit [Ping timeout: 265 seconds]
yfeldblum has joined #jruby
<GitHub74>
[jruby] headius closed issue #3304: zone_spec.rb fails SPEC test "Time#zone returns UTC when called on a UTC time" http://git.io/vGXQb
<GitHub90>
[jruby] headius closed issue #3303: strftime_spec.rb SPEC test fails in "returns the timezone with %Z" http://git.io/vGXVh
<headius>
xardion: good to know...do you mean you also deploy on 8?
<GitHub7>
[jruby] headius pushed 1 new commit to jruby-1_7: http://git.io/vG7kt
<GitHub7>
jruby/jruby-1_7 9e2d6dc Charles Oliver Nutter: Do not process arguments twice. Fixes #3294.
<GitHub179>
[jruby] headius pushed 1 new commit to master: http://git.io/vG7k2
<GitHub179>
jruby/master 1cbb8d6 Charles Oliver Nutter: Merge remote-tracking branch 'origin/jruby-1_7'
<GitHub198>
[jruby] headius closed issue #3294: processArguments is run twice http://git.io/vGt1M
camlow325 has quit [Remote host closed the connection]
yfeldblu_ has joined #jruby
benlovell has joined #jruby
camlow325 has joined #jruby
shellac has quit [Quit: Computer has gone to sleep.]
yfeldblum has quit [Ping timeout: 256 seconds]
camlow325 has quit [Remote host closed the connection]
<nirvdrum>
readlink has been unhandled on Windows in jnr-posix since 2008.
<nirvdrum>
I don't know if it was ever implemented in Ruby.
<nirvdrum>
Err, in Java.
<nirvdrum>
It's a weird one. NTFS has had joinpoints for a long time. I think Windows 8.1 or 10 added symlinks. I'm not sure what the difference actually is.
pitr-ch has quit [Ping timeout: 260 seconds]
subbu|away is now known as subbu
shellac has joined #jruby
lance|afk is now known as lanceball
benlovell has joined #jruby
benlovell has quit [Ping timeout: 272 seconds]
benlovell has joined #jruby
camlow325 has quit [Remote host closed the connection]
ivan\ has quit [Quit: ERC Version 5.3 (IRC client for Emacs)]
camlow325 has joined #jruby
camlow325 has quit [Remote host closed the connection]
camlow325 has joined #jruby
camlow325 has quit [Remote host closed the connection]
camlow325 has joined #jruby
<GitHub180>
[jruby] headius pushed 1 new commit to jruby-1_7: http://git.io/vG7aU
<GitHub180>
jruby/jruby-1_7 2797b9b Charles Oliver Nutter: Mark File.readlink as not impl on Windows. Sorta fixes #3287....
shellac has quit [Ping timeout: 240 seconds]
havenwood has joined #jruby
ivan\ has joined #jruby
<GitHub80>
[jruby] headius pushed 1 new commit to master: http://git.io/vG7a8
<GitHub80>
jruby/master 70c6156 Charles Oliver Nutter: Merge remote-tracking branch 'origin/jruby-1_7'
<GitHub192>
[jruby] headius closed issue #3287: Cannot resolve or detect symlinks on Windows http://git.io/vsxFG
bbrowning is now known as bbrowning_away
ivan\ has quit [Remote host closed the connection]
<GitHub50>
[jruby] headius closed issue #3289: Jruby 9.0.0.0: ActionView::Template::Error (can't add a new key into hash during iteration) http://git.io/vshJM
<nirvdrum>
headius: Do you or enebo want to try to add AppVeyor for JNR? Or do you want to temporarily make me an admin and I'll work through it?
<headius>
I'm happy to add you as an admin
<enebo>
headius: see if I am even on appveyor while you are at it
<nirvdrum>
Cool. I think I have the config correct, but I can't add the necessary GitHub hooks.
<headius>
eh?
<headius>
nirvdrum: ahh makes sense
<nirvdrum>
Naturally we don't pass everything on Windows. But it's a starting point.
<headius>
nirvdrum: you're in
<headius>
enebo: on appveyor? I don't know what you mean
dinfuehr has joined #jruby
<enebo>
headius: how did you add him?
ivan\ has joined #jruby
<enebo>
headius: can you check to see if he is an admin now?
<nirvdrum>
Thanks.
<enebo>
headius: transitively can you see if I am one?
<headius>
enebo: on github
<headius>
not on appveyor
<enebo>
oh heh
<enebo>
headius: ok I misread that and I was fairly certain I had never used appveyor :)
<headius>
and you are indeed an owner on github :-)
<nirvdrum>
headius, enebo: I'm still working out the AppVeyor team stuff, but I think if you guys have an account I can add you an admins to the one I've set up for JNR.
<nirvdrum>
Heh, actually I can set up the jnr/Owners team. That'll be easier.
<nirvdrum>
headius, enebo: It may make sense to set up AppVeyor for JRuby in general. I don't think we necessarily want as many configurations as we have in Travis, but it could help catch Windows-specific issues earlier on.
<headius>
it would absolutely make sense
skade has joined #jruby
<headius>
excellent...I have wanted to play with this but never have
benlovell has quit [Ping timeout: 264 seconds]
<nirvdrum>
Well, I'm clocking out for vacation soon :-) But I can look at that when I get back.
<nirvdrum>
Figuring out the build steps was kind of annoying. Their docs aren't super great.
<nirvdrum>
My preference would have been for CloudBees to do this so we don't have yet another CI provider in the mix. But such is life.
<nirvdrum>
Or Travis doing it for that matter.
<nirvdrum>
They don't seem to go back to Java 6. I think we could install it on our own, but I've punted on that for now.
<nirvdrum>
Technically jnr-posix needs to support Java 6 because it's still used by JRuby 1.7.
havenwood has joined #jruby
<GitHub168>
[jruby] headius pushed 1 new commit to jruby-1_7: http://git.io/vG7D9
<GitHub168>
jruby/jruby-1_7 4bee414 Charles Oliver Nutter: Update to jnr-ffi-2.0.4-SNAPSHOT for #3258....
<headius>
yeah, bees would be right but I don't think our account has access to Windows (if they even have Windows)
<headius>
damn, transposed digits
lanceball is now known as lance|afk
enebo has quit [Quit: enebo]
<GitHub178>
[jruby] headius pushed 1 new commit to master: http://git.io/vG7y6
<GitHub178>
jruby/master 34c2316 Charles Oliver Nutter: Update to jnr-ffi-2.0.4-SNAPSHOT for #3285....
<headius>
settin 'em up and knockin 'em down
skade has quit [Quit: Computer has gone to sleep.]
<GitHub46>
[jruby] headius closed issue #3285: "NotImplementedError: waitpid unsupported or native support failed to load" with 9.0.0.0 on Ubuntu if build tools not installed http://git.io/vsNmm
yfeldblu_ has quit [Remote host closed the connection]
yfeldblum has joined #jruby
havenn has joined #jruby
havenwood has quit [Ping timeout: 265 seconds]
_djbkd has quit [Remote host closed the connection]
Aethenelle has joined #jruby
_djbkd has joined #jruby
samphippen has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<oblutak>
issue #3254 is closed and included in 1.7.22, but it didn't make it in 9.0.1.0. Should I open a new issue for 9k or I need to wait for the 'real' fix in jnr-posix?
<GitHub123>
[jruby] eregon commented on commit 7a99f1c: The edit in place might work to some extent if the branch was on the shared jruby/jruby repo. http://git.io/vG5vV
<oblutak>
headius: I have to run, but I'll open a new issue tomorrow. I am fairly new to git so I'm probably missing something. I'm just looking at code via web interface, but I don't see the Solaris check on the RubyFile.java version tagged with 9.0.1.0.
skade has joined #jruby
oblutak has quit [Ping timeout: 246 seconds]
dinfuehr has joined #jruby
dinfuehr has quit [Ping timeout: 244 seconds]
tcrawley is now known as tcrawley-away
skade has quit [Quit: Computer has gone to sleep.]