<headius>
then the next build with unrelated truffle stuff produced this:
<headius>
Gem::InstallError: invalid gem: package is corrupt, exception while verifying: undefined method `size' for nil:NilClass (NoMethodError) in /home/travis/.m2/repository/rubygems/ruby-maven-libs/3.1.1/ruby-maven-libs-3.1.1.gem
<headius>
error: error reading /home/travis/.m2/repository/org/jruby/jcodings/jcodings/1.0.13-SNAPSHOT/jcodings-1.0.13-SNAPSHOT.jar; error in opening zip file
<headius>
something's really screwed up
<headius>
I'm going to kill all builds and reset maven caches again
<mkristian>
mkristian, tesla does download some jruby not sure which 1.7.12 or so when it is not in the local maven repo. what I really find surprising is that all the jdk8 builds pass this point.
<headius>
yes
<headius>
that's really strange
<mkristian>
how do you reset the maven caches ?
rsim has quit [Quit: Leaving.]
<headius>
travis settings for jruby/jruby
<mkristian>
ok
<mkristian>
are the jenskin build doing as usual or are there similar problems - beside the missing snakeyaml.jar ?
<headius>
they're all green
<headius>
:-)
<headius>
that could indicate there's something seriously broken on travis
<headius>
sorry I assumed it was your doing, but those first couple builds obviously pointed the wrong way :-)
<mkristian>
no problem. I did see the first 2 reds and they did not look related to me. so I did not care that my commit was red as well. I did not notice the that things got worth by hour yesterday
<mkristian>
until rtyler pointed towards it
<mkristian>
it seems to go beyond the jruby-libs module now
<headius>
yeah
<headius>
I have a theory
pgokeeffe has joined #jruby
<headius>
I tried to reset the maven caches the other day when a few builds were cgoing
<headius>
maybe it got corrupted and just went downhil
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] mkristian deleted test-build at a07a15d: http://git.io/vfs6I
JRubyGithub has left #jruby [#jruby]
<headius>
well it looks ok now
<mkristian>
but good to know that I also can delete the caches whenever I see something pointing to corrupted artifact
nirvdrum has quit [Ping timeout: 264 seconds]
<mkristian>
so the snakeyaml.jar is really missing the dist tarball. looking now
<JRubyGithub>
[jruby] trejkaz opened issue #2865: Ambiguous method warning when there does not appear to be any ambiguity http://git.io/vfs1g
JRubyGithub has left #jruby [#jruby]
<headius>
mkristian: it's starting to fall apart again
<headius>
Failed to execute goal de.saumya.mojo:gem-maven-plugin:1.0.7:initialize (default) on project jruby-tests: Execution default of goal de.saumya.mojo:gem-maven-plugin:1.0.7:initialize failed: A required class was missing while executing de.saumya.mojo:gem-maven-plugin:1.0.7:initialize: org/apache/tools/ant/BuildListener
<headius>
I think there's something wrong with their maven caching
<headius>
we're going to have to turn it off and report it
<headius>
I will let this build finish and see how bad it gets
<mkristian>
this is the jar which contains the missing class. concluding this jar is corrupt
subbu has quit [Ping timeout: 245 seconds]
a5i has quit [Quit: Connection closed for inactivity]
<mkristian>
jruby-stdlib.jar which is build inside ./lib does already miss the snakeyaml.jar on jenskin but not with me locally.
<headius>
hmm
<headius>
which jar is corrupt?
<mkristian>
jenskin is really nice that I can browse the working directory of the last build
<mkristian>
the ant-19.4.jar
<mkristian>
is corrupt
<headius>
right
<headius>
it was a jruby class previously though
<headius>
I think it's something screwy with how they're doing the maven caching
<mkristian>
yes, feels like some currency issue - just thinking
<mkristian>
I see reason for the snakeyaml.jar on jenskin and how to fix it :)
<headius>
yeah concurrency was my thought too
<headius>
actually I'm just going to restart this build...if theory holds, it should mostly fail
<headius>
mkristian: what did you find?
<mkristian>
headius, snakeyaml.jar gets copied into the right place AFTER the jruby-stdlib.jar is packed
<headius>
oh!
<headius>
and locally unless you scrub your env it's already there
<headius>
I scrubbed and reproduced the same problem
<headius>
this is probably the source of the other snakeyaml weirdness
<mkristian>
and jruby-dist uses jruby-stdlib.jar. locally I still looking why it isi there. probably this is cause of other snakeyaml issues
<headius>
this day is turning out a lot better
<mkristian>
hope so . . . and my time is limited since I am flying back to london and then munich this evening (IST)
yfeldblum has joined #jruby
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<mkristian>
locally it indeed says: snakeyaml-1.14.jar already exists in destination. will fix the order now
iamjarvo has joined #jruby
yfeldblu_ has quit [Ping timeout: 245 seconds]
<headius>
ok
<headius>
mkristian: thank you for looking into it
<headius>
new build is looking the same
<mkristian>
new build on travis ?
<headius>
mkristian: just re-ran most recent
<headius>
one of the failures is this: size of maven/jruby-dist/target/jruby-dist-9.0.0.0-SNAPSHOT-src.zip expected smaller then 13*1024*1024 but got 13639091
iamjarvo has quit [Client Quit]
<mkristian>
I think it just grew bigger. granularity is 1MB
pgokeeffe has quit [Quit: pgokeeffe]
<mkristian>
is is meant to cache those explosions of distribution artifacts we saw in the past
<headius>
right
<headius>
ok, I'll bump it
<headius>
mkristian: ok, I give up, where is that verification?
<mkristian>
in test/check.sh
<headius>
ahh right
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] mkristian pushed 3 new commits to master: http://git.io/vfsQu
<JRubyGithub>
jruby/master 336ed58 Christian Meier: copy snakeyaml.jar before packing the jar [skip ci]
<JRubyGithub>
jruby/master 4a947d4 Christian Meier: follow the changes from core/pom.xml into core/pom.rb
<JRubyGithub>
jruby/master 2bea720 Christian Meier: Merge branch 'master' of github.com:jruby/jruby
JRubyGithub has left #jruby [#jruby]
<mkristian>
I just pushed the snakeyaml fix - could you trigger a jenskin build ?
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<chrisseaton>
no I mean the way you bulk untag RubySpec specs - that's not this is it?
havenwood has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] bjfish pushed 1 new commit to master: http://git.io/vfn37
<JRubyGithub>
jruby/master d5f5320 Brandon Fish: [Truffle] Updating test index for passing tests.
JRubyGithub has left #jruby [#jruby]
<bjfish2>
chrisseaton there are two tools in the in the tools/truffle directory there are some comments at the top of each
<chrisseaton>
bjfish2: ah that was it - thanks
<chrisseaton>
bjfish2: we've passed 70% btw as well
<bjfish2>
chrisseaton yes, i saw thats great, +20% in less than 2 months i think
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
<chrisseaton>
yeah, and we have more lined up with the pack specs and work on IO
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
kares has joined #jruby
iamjarvo has joined #jruby
<bjfish2>
chrisseaton BTW the ENV tests pass in isolation with the bulk untaggers but then when you go to run them in a test suite they mess up other tests because the ENV tests clear all the ENV variables currently
<bjfish2>
chrisseaton I usually run the run_each_core_dir.rb tool because it finishes in a reasonable amount of time
<chrisseaton>
I can just modify files = to make it run on a specific subdirectory can't I?
iamjarvo has quit [Ping timeout: 256 seconds]
<chrisseaton>
Got it working, thanks
vtunka has joined #jruby
<bjfish2>
chrisseaton okay good, I usually do something like this: files.select! { |f| f.include?("io") }; puts "Files: #{files}" , but changing the glob would work too
lance|afk is now known as lanceball
enebo has joined #jruby
mcclurmc has joined #jruby
mcclurmc_ has quit [Ping timeout: 255 seconds]
e_dub has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] eregon pushed 1 new commit to master: http://git.io/vfngL
<JRubyGithub>
jruby/master 5ccc678 Benoit Daloze: [Truffle] Rewrite a good part of RubyFiber and RubyThread....
JRubyGithub has left #jruby [#jruby]
<headius>
eregon, chrisseaton: doesn't truffle or graal have anything for continuations yet?
<headius>
that would be a huge motivator
iamjarvo has joined #jruby
<headius>
well not full continuations, just coroutines
iamjarvo has quit [Max SendQ exceeded]
<eregon>
headius: I plan to invetigate the coroutine support
<eregon>
yeah, continuation might be possible but it's very invasive in the codebase and anyway it's sort of deprecated nowadays
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
<eregon>
but first, I need to fix some of RubyFiber and thread :)
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
<headius>
they aren't pursuing coroutines anymore?
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
subbu has joined #jruby
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
skade has joined #jruby
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
e_dub has quit [Quit: ZZZzzz…]
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
<eregon>
headius: what do you mean?
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] eregon pushed 2 new commits to master: http://git.io/vfno2
<JRubyGithub>
jruby/master 8fcbcb8 Brandon Fish: [Truffle] Add MRI test profile to pom files.
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] chrisseaton pushed 5 new commits to truffle-io: http://git.io/vfnPZ
<JRubyGithub>
jruby/truffle-io e08fd5f Chris Seaton: Merge branch 'master' into truffle-io...
<JRubyGithub>
jruby/truffle-io 12ad44e Chris Seaton: [Truffle] Allow module_function in a class when loading the core.
<JRubyGithub>
jruby/truffle-io e8835c0 Chris Seaton: [Truffle] fix the write primitive.
JRubyGithub has left #jruby [#jruby]
ypasmk_ has quit [Remote host closed the connection]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] chrisseaton opened pull request #2866: [Truffle] New IO implementation (master...truffle-io) http://git.io/vfnM2
JRubyGithub has left #jruby [#jruby]
<headius>
chrisseaton: I hate to say this but the IO work I did recently is going to remain a lot closer to MRI compat than the rbx impls
<headius>
there's a lot of stuff missing from them
<chrisseaton>
headius: yeah - it's all about getting stuff working quickly - we are prepared to re-implement this in the future if required - but from a position of having most stuff working, rather than working from nothing
<chrisseaton>
headius: we need Socket soon, and this is the foundation for that
arturaz has quit [Ping timeout: 244 seconds]
<headius>
that's fine, just wanted to give you a heads up
<chrisseaton>
what were the key issues though?
<headius>
lots and lots of encoding stuff
<chrisseaton>
I think cheald or someone was also planning to reimplement the Rubinius code for the same reasons you are concerned about
<chrisseaton>
thanks to bjfish we are almost running MRI tests as well now
<headius>
yeah that's very good
<headius>
also +1 on prioritizing coro work...I think it could be as big a game-changer as invokedynamic was supposed to be :-)
<headius>
(and generally is, I'll be fair)
<chrisseaton>
the problem with coros is that it's an full on modification to the JDK - way bigger than Graal
<chrisseaton>
you probably need to pester product on that one
<headius>
yeah I know
<headius>
and I do
<headius>
I just don't have a golden hammer to beat them with
iamjarvo has joined #jruby
nirvdrum has joined #jruby
rsim has quit [Quit: Leaving.]
<nirvdrum>
enebo: I think the lexer_source_encoding branch fixes the problems we were discussing. I was pushing stuff to coordinate between Windows and Linux, so the commits should probably just be squashed.
<nirvdrum>
A couple other interesting things came out of it though.
<nirvdrum>
E.g., I set up that Japanese locale on my Linux machine and the charset name ends up being x-euc-jp-linux. That can't be looked up in jcodings, so we fall back to US-ASCII.
<lopex>
nirvdrum: also, some expert should look how our encodings relate to those in java wrt features
<nirvdrum>
lopex: Alas, I think there's only 4 of us that really understand any of this :-/
<lopex>
ah
<lopex>
nirvdrum: things like supported properties and the like
<nirvdrum>
And really, the 3 of you understand it far better than I.
pierreatkillbill has quit [Ping timeout: 240 seconds]
rsim has joined #jruby
mrmargol_ has joined #jruby
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<chrisseaton>
headius: are the IO encoding problems represented in RubySpec? You'd be free to put them in now, if there were disagreements in the past, of course
vtunka has quit [Quit: Leaving]
benlovell has quit [Ping timeout: 276 seconds]
mrmargolis has quit [Ping timeout: 248 seconds]
benlovell has joined #jruby
rcvalle has joined #jruby
havenwood has quit [Remote host closed the connection]
baroquebobcat has joined #jruby
rsim has quit [Quit: Leaving.]
pitr-ch has quit [Quit: My Mac Mini has gone to sleep. ZZZzzz…]
<headius>
lopex: all your changes on joni master are safe?
<headius>
we hadn't done a snapshot in a while so we weren't testing with them
<headius>
perhaps I should force a snapshot and re-run a green build
<lopex>
headius: I would tentatively say yes
<headius>
ok
<headius>
looks like mostly zero-sum
<lopex>
and new features are onigmo test suite covered
<lopex>
headius: looking at the changes ?
<headius>
just a couple
<headius>
I trust you :-)
<lopex>
yeah, either zero sum or new features that should affect existing code
<headius>
ok, joni snapshot is updated...I'll wait to release until a build completes
<lopex>
I stopped before major more invasive changes
pierreatkillbill has joined #jruby
shellac has quit [Ping timeout: 240 seconds]
<lopex>
bbrowning_away: also, those changes will affect javascript syntaxt for joni
<lopex>
flavorjones: oh, btw, how do I emit via to_xml and have html like <script..></script> closing tags ?
rsim has joined #jruby
<lopex>
headius: doh, I forgot to respond to that facebook mail :(
<lopex>
I'll have to make big excuses
x1337807x has joined #jruby
<lopex>
oh NO_EMPTY_TAGS
zorak8 has joined #jruby
bbrowning_away is now known as bbrowning
rsim has quit [Quit: Leaving.]
<bbrowning>
lopex: which changes?
<bbrowning>
the dynjs project is the only thing that consumes the javascript syntax iirc
<bbrowning>
hey torquebox 4 tests are all green on jruby 9.0.0.0-dev!
<headius>
wtf
<headius>
if that doesn't say ship it I don't know what will
<bbrowning>
it means I need to ship another TB4 release, at least :)
<lopex>
does it also compile ?
* lopex
runs
<headius>
:-)
<lopex>
bbrowning: changes related to ascii char type modifiers
<bbrowning>
ahh
lanceball is now known as lance|afk
<lopex>
bbrowning: all that code is no more in onigmo
<bbrowning>
I'll make a note to take a look at it next time we need to touch regexps in dynjs
<lopex>
bbrowning: well, we'll have to choose different places, OR it wont be any issue after catching with onigmo changes
mje113__ has joined #jruby
<gazay>
hello!
jeremyevans has quit [Ping timeout: 256 seconds]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] chrisseaton pushed 1 new commit to master: http://git.io/vfcdP
<JRubyGithub>
jruby/master 5385f46 Chris Seaton: [Truffle] Proc#yield
JRubyGithub has left #jruby [#jruby]
<gazay>
headius: I plan to start investigating failing tests in AS. I think most of them I will not be able to fix and send PR by myself as I’m not good in Java. What would be better way to give you information of what I found – create issue in jruby repo or just create gist and send it here?
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] chrisseaton pushed 1 new commit to master: http://git.io/vfcFR
<JRubyGithub>
jruby/master bae615c Chris Seaton: [Truffle] Try moving Kernel#loop to Rubinius.
JRubyGithub has left #jruby [#jruby]
mrmargolis has joined #jruby
rsim has joined #jruby
baroquebobcat has quit [Quit: baroquebobcat]
<headius>
gazay: issues for sure, and if you can give us shortened reproductions that would save 90% of the work
mrmargol_ has quit [Ping timeout: 248 seconds]
<headius>
almost all of these will be simple fixes, but we'd spend a lot of time hunting down reproductions
<gazay>
sure! btw I’ve pulled from rails master (it was pretty old locally – from middle of march) and now there ~100 failing tests =(
<gazay>
in AS
<headius>
I'm sure many are failing for the same reasons
<headius>
and there's a big block of failures in a row that are probably all the same thing
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] bjfish pushed 1 new commit to master: http://git.io/vfcxf
<headius>
it was entertaining (in retrospect) to watch the builds completely fall apart
<headius>
and we are somewhat unique in how many jobs we run each build, so we get a ton of instant concurrency
djellemah_ has quit [Ping timeout: 265 seconds]
<bbrowning>
headius: how was the corrupt jar thing manifesting itself?
<headius>
all sorts of weird shit
<headius>
classes being missing, errors about jars not being openable
<headius>
it could be any jar we used in maven so the failures were all over the place and all confusing
<bbrowning>
huh I think I'm hitting the same thing with my TB build - it keeps having problems unzipping a wildfly-dist.jar
<headius>
that's it
<headius>
we had cases like that too
<headius>
try turning off cache or resetting it for a build
<bbrowning>
yeah I am using the cache
<headius>
ultimately it was process of elimination...none of it made sense, but corrupt jars was the only plausible explanation for all of them
<bbrowning>
good catch - I've retried a build 3x now and it keeps failing when unzipping this jar, so I'm sure that's it
<bbrowning>
it keeps getting the corrupted version from cache
brightball has joined #jruby
<headius>
yup exactly
<headius>
and we'd see one bad build cascade into multiple the next run
<headius>
as more jars got corrupted
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] gazay opened issue #2867: Date#next_year difference behvior with MRI http://git.io/vfCOy
JRubyGithub has left #jruby [#jruby]
pitr-ch has joined #jruby
<headius>
gazay: it has begun!
<headius>
now you're going to get people excited about fixing stuff :-)
rsim has joined #jruby
<gazay>
headius: =) is format of issue ok?
<headius>
if the example gets much larger you should just make it a runnable Ruby script rather than an IRB session, but this is fine for one-liners
<headius>
ruby script = one code section for the script, one for the output, and one for comparative output on MRI
<gazay>
got it
<headius>
just so we can easily run larger examples
lance|afk is now known as lanceball
<chrisseaton>
Does anyone know why I might get a JONI 'empty range in char class' ValueException for /[\u{10000}-\u{EFFFF}]/u JRuby handles it fine but Truffle (also using JONI) doesn't
<headius>
bad encoding fed to joni?
<headius>
such that it doesn't recognize that range as valid characters
<headius>
chrisseaton: one succeeded so I restarted that whole build
<headius>
actually I restarted 0574
<headius>
the next one I think
lxsameer has quit [Quit: Leaving]
pierreatkillbill has quit [Ping timeout: 256 seconds]
baroquebobcat has quit [Quit: baroquebobcat]
dinfuehr has joined #jruby
baroquebobcat has joined #jruby
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
dinfuehr has quit [Ping timeout: 256 seconds]
elia has joined #jruby
<gazay>
headius: I’m going to sleep (its 5am here) and will continue tomorrow morning. I’m stuck with strange behavior of Module#attr_* methods in several rails’ tests and in jruby irb. I didn’t create issue because I’m not sure that I understood the problem there. Please take a look when you have a time. https://gist.github.com/gazay/a51dee0bd6bcafb65dfd
subbu has joined #jruby
<headius>
gazay: ok good progress so far :-)
<headius>
thanks for your help!
<headius>
chrisseaton: looking better now
<headius>
enebo: joni, jcodings, bytelist, jnr-posix, jnr-enxio, and jnr-unixsocket have been released...just waiting on propagation to push pom update
<headius>
I will update 1.7 too
<headius>
enebo: there are a bunch of deps in 1.7 that are not latest release, but I'm not sure what our policy is on that sort of thing
<headius>
I suppose there are likely unreported bugs that get fixed by updating
mrmargol_ has joined #jruby
mrmargolis has quit [Ping timeout: 245 seconds]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] chrisseaton pushed 5 new commits to master: http://git.io/vfW6Y
<JRubyGithub>
jruby/master e6af2c6 Chris Seaton: [Truffle] finish Array#hash
<JRubyGithub>
jruby/master 4943ce2 Chris Seaton: [Truffle] Array#initialize will call a private to_ary
<JRubyGithub>
jruby/master c8ef4c3 Chris Seaton: [Truffle] Tag failing Fixnum#^ spec