These are chat archives for jruby-gradle/jruby-gradle-plugin

2nd
Jun 2015
R. Tyler Croy
@rtyler
Jun 02 2015 12:45
moin
Christian Meier
@mkristian
Jun 02 2015 12:47
moin
R. Tyler Croy
@rtyler
Jun 02 2015 12:50
@ysb33r you ever see behavior where the classpath gets messed with in multiproject builds for plugins?
I just noticed this project I have applys the "shadow" plugin, but it never adds it to the classpath and it's not a transitive dependency of the plugins applied to that sub-project build
it is however a transitive of the parent project's plugins
rtyler @rtyler boggles
Schalk W. Cronjé
@ysb33r
Jun 02 2015 13:44
I depends how you have set up your dependencies. It could be because your are using compile project(’:myParent’) or you’ve set it up in the root project’s build.gradle
R. Tyler Croy
@rtyler
Jun 02 2015 13:47
but for the buildscript dependencies does that apply?
Schalk W. Cronjé
@ysb33r
Jun 02 2015 13:54
buildscript in root project can propagate to subproject I think, but I need to test it.
R. Tyler Croy
@rtyler
Jun 02 2015 13:54
as far as I can tell, the application of the plugin's DSL method does not propagate
but the classpath is shared
(from my observations)
Schalk W. Cronjé
@ysb33r
Jun 02 2015 13:54
it is a nasty fella, this one called buildscript. It tends to screw with your mind.
ysb33r @ysb33r says hello from Gr8Conf
R. Tyler Croy
@rtyler
Jun 02 2015 13:55
o/
Schalk W. Cronjé
@ysb33r
Jun 02 2015 14:22
Just had an idea - maybe you guys can use lazybones to bootstrap projects very quickly
Even better than making them copy-paste a Gradle example
R. Tyler Croy
@rtyler
Jun 02 2015 14:23
@ysb33r in what context?
Schalk W. Cronjé
@ysb33r
Jun 02 2015 14:26
Like you told me the other day, when you were struggling with gradleTest. Maybe lazybones could be a way to bootstrap jruby+gradle projects too...
I’ve never thought about it until today
I’m just not sure how lazybones deal with a local repo of templates
R. Tyler Croy
@rtyler
Jun 02 2015 14:32
I was struggling with gradleTest for other reasons, but lazybones is an interesting idea for "user bootstrap" for jruby-gradle in general IMO
Schalk W. Cronjé
@ysb33r
Jun 02 2015 15:15
It would be sweet to put one together for jrubyconf.eu ...
R. Tyler Croy
@rtyler
Jun 02 2015 15:15
fucking awesome, Gradleware broke permalinks to forums.gradle.org posts in the new forums
Schalk W. Cronjé
@ysb33r
Jun 02 2015 15:15
they should redirect
R. Tyler Croy
@rtyler
Jun 02 2015 15:15
great piece of technology, but they really realy suck at running a project
@ysb33r will you file a ticket in the base plugin so we don't forget?
Schalk W. Cronjé
@ysb33r
Jun 02 2015 15:16
every link I foudn so far redirected to the correct one on the new forum.
R. Tyler Croy
@rtyler
Jun 02 2015 15:20
the ones I've put in internal bug trackers haven't :(
this dependency resolution just bit me in the ass again with JRuby Gradle
so infuriating
@mkristian already fixed it, but it's not released, it's such incorrect behavior IMO
R. Tyler Croy
@rtyler
Jun 02 2015 15:44
@mkristian looks like we might have a bug with the classloader stuff or I'm being completely stupid :)
RuntimeError:

        you might need to reinstall the gem which depends on the missing jar or in case there is Jars.lock then JARS_RESOLVE=true will install the missing jars

no such file to load -- org/apache/kafka/kafka_2.10/ /kafka_2.10- -0.8.1.1 (LoadError)
          do_require at uri:classloader:/gems/jar-dependencies-0.1.14/lib/jar_dependencies.rb:254
org/apache/kafka/kafka_2.10/0.8.1.1/kafka_2.10-0.8.1.1.jar is in the artifact and the jrubydirs are there:
org/apache/kafka/
org/apache/kafka/.jrubydir
org/apache/kafka/kafka_2.10/
org/apache/kafka/kafka_2.10/.jrubydir
org/apache/kafka/kafka_2.10/0.8.1.1/
org/apache/kafka/kafka_2.10/0.8.1.1/.jrubydir
those spaces are concerning to me
Christian Meier
@mkristian
Jun 02 2015 16:08
@rtyler I did see this before: org/apache/kafka/kafka_2.10/ /kafka_2.10- -0.8.1
how does the Jars.lock look like ? that is where jar-dependencies loads all jars from
I thought I fixed it
R. Tyler Croy
@rtyler
Jun 02 2015 16:10
let me gist
Christian Meier
@mkristian
Jun 02 2015 16:11
yes, that is the same kind as before the fix I just mentioned
R. Tyler Croy
@rtyler
Jun 02 2015 16:11
what fix? something in the unreleased code inthe base plugin?
Christian Meier
@mkristian
Jun 02 2015 16:11
just looking for it
R. Tyler Croy
@rtyler
Jun 02 2015 16:12
danke
Christian Meier
@mkristian
Jun 02 2015 16:13
it is on base plugin bascially the second last commit: 9aa28c9fe9881fb3c5b45f2b2a9454b8b1e151c1
and it does not look as it is released
R. Tyler Croy
@rtyler
Jun 02 2015 16:14
guess I'm cutting releases then :)
Christian Meier
@mkristian
Jun 02 2015 16:14
@ysb33r @rtyler any chance to get the last PR of mine reviewed and released ?
R. Tyler Croy
@rtyler
Jun 02 2015 16:15
for the jar plugin?
Christian Meier
@mkristian
Jun 02 2015 16:15
yes
R. Tyler Croy
@rtyler
Jun 02 2015 16:19
cutting 0.2.0 of the base plugin assuming everything passes
we need to get gradletests for all these suckers :)
Schalk W. Cronjé
@ysb33r
Jun 02 2015 16:20
you can write them … then I get better feedback on the plugin :smile:
R. Tyler Croy
@rtyler
Jun 02 2015 16:21
@mkristian I published 0.2.0, will you update the changelog properly for yor changes that went into it?
Christian Meier
@mkristian
Jun 02 2015 16:22
ok
rtyler @rtyler waits impatiently for the plugin portal to update
Schalk W. Cronjé
@ysb33r
Jun 02 2015 16:32
Yesm please, I was just about to ask for a CHANGELOG as I cannot keep up with changes anymore
Christian Meier
@mkristian
Jun 02 2015 16:35
will do next - just give me a few minutes
R. Tyler Croy
@rtyler
Jun 02 2015 16:36
heh, now the version resolution at least fails
instead of being stupid and optimistically wrong
@mkristian sent you an internal gist on internal IM
Christian Meier
@mkristian
Jun 02 2015 16:37
if you change the order of dependencies things could pass - I know it is not perfect but hopefully one step closer
R. Tyler Croy
@rtyler
Jun 02 2015 16:37
it is definitely at least surfacing a somewhat useful error :)
I'm not sure it's correct in some of the behaviors though, but i'm less familiar with bundler than may be required
Christian Meier
@mkristian
Jun 02 2015 16:38
yes but looks like many errors :(
R. Tyler Croy
@rtyler
Jun 02 2015 16:38
> there is no overlap for [1.4,1.99999] and [0,)
that for example
seems erroneous
while this: > there is no overlap for [0.3,) and [0.4.2,) is 100% correct
er
sorry, also not correct
> there is no overlap for [1.4.1,1.4.99999] and [1.3.0,1.3.99999] that's what I meant to paste
Christian Meier
@mkristian
Jun 02 2015 16:39
ahh - buh - empty open brackets: [0.3,)
R. Tyler Croy
@rtyler
Jun 02 2015 16:40
huh?
I don't parse maven dependency eccentricities well :P
Christian Meier
@mkristian
Jun 02 2015 16:40
the last one has no overlap: [1.4.1,1.4.99999] and [1.3.0,1.3.99999]
R. Tyler Croy
@rtyler
Jun 02 2015 16:41
yeah, that is what I meant to say is a correct error to me
whereas the previous two examples seem like the gem resolver needs to be updated
Christian Meier
@mkristian
Jun 02 2015 16:41
what shall I do there ?
R. Tyler Croy
@rtyler
Jun 02 2015 16:41
well, am I crazy? is this expected behavior? that you can answer first :P
perhaps I could use gradleTest as a means of introducing regression test cases for you if these are in fact bugs
Christian Meier
@mkristian
Jun 02 2015 16:42
yes, [0.3,) needs to allow all versions up to infinity
well if I have no overlap then I can do the gradle thing: pick the newest. BUT it just could resolve itself once all the other cases work
did you say there is Gemfile.lock for this project as well ?
R. Tyler Croy
@rtyler
Jun 02 2015 16:44
there is
Christian Meier
@mkristian
Jun 02 2015 16:45
coud you gist me this one as well
R. Tyler Croy
@rtyler
Jun 02 2015 16:54
@ysb33r apologies for some of the internal back-channel discussion, this project I'm working with is internal and has tons of internal dependencies :/
we'll get some regression tests into the public project as soon as we can
R. Tyler Croy
@rtyler
Jun 02 2015 17:15
hooray, plugin portal updated
R. Tyler Croy
@rtyler
Jun 02 2015 17:48
#124 filed to capture these issues with lots of REDACTED :D
Schalk W. Cronjé
@ysb33r
Jun 02 2015 17:58
I’m at hackergarten - just ignoring you :)
R. Tyler Croy
@rtyler
Jun 02 2015 18:03
hackergarten?
rtyler @rtyler feels ignored
Christian Meier
@mkristian
Jun 02 2015 18:08
he told you to ignore you
R. Tyler Croy
@rtyler
Jun 02 2015 18:10
+--- rubygems:rspec-core:2.99.+ -> 3.2.3
gahhhhhhhhhhhhhhhh that's so fucking aggrevating
Gradle, this is so wrong
Christian Meier
@mkristian
Jun 02 2015 18:11
this dependency resolution of gradle is the only bad thing I encountered so far. everything else seems to really well thought out
Christian Meier
@mkristian
Jun 02 2015 19:43
PR for two rather small changes to GemVersion jruby-gradle/jruby-gradle-plugin#125
R. Tyler Croy
@rtyler
Jun 02 2015 19:43
woohoo!
@mkristian this should resolve #124 correct?
Christian Meier
@mkristian
Jun 02 2015 19:48
I would say so. even if the one conflict does has resolution if you look at all possibilities. but with callback the plugin gets from gradle this can not be resolved
maybe there is some other way to implement this feature but the current road can not achieve more
R. Tyler Croy
@rtyler
Jun 02 2015 20:05
cutting a new release
Christian Meier
@mkristian
Jun 02 2015 20:05
thanx
R. Tyler Croy
@rtyler
Jun 02 2015 20:13
@mkristian now there's two changelog entries we need :)
R. Tyler Croy
@rtyler
Jun 02 2015 20:19
@mkristian integration testing now
@mkristian I still don't think this is entirely correct though
I'm trying to isolate the issue, but it looks like the behavior around this rack dependency is the same now as it was pre-0.2.0