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

22nd
Apr 2015
Christian Meier
@mkristian
Apr 22 2015 03:27
I am about to add a testcase for this branch. each PR is on its own branch - like #108 is on 'generate-dir-info' and I will add a testcase there as well (or better just another expect or so)
and for jruby-gradle-jar-plugin it is the 'jruby-mains' branch but really depends on #110
Christian Meier
@mkristian
Apr 22 2015 10:29
each PR is in its own branch. #110 has a small integration test now
#108 does already have a minimal test. to combine the two jruby execution into one is still pending but maybe it is already worth merging
as far I can say both PRs do not really interfere with with existing features - unlike the PR on jruby-gradle-jar-plugin which does changes things quite a bit
Schalk W. Cronjé
@ysb33r
Apr 22 2015 10:35
which ones do you want me to look at first? these ones on ruby-gradle-base itself?
fn ‘ell, is there no way to just download the diff from github ???
Ahhhh, adding .diff to the end of the URL does the trick
Christian Meier
@mkristian
Apr 22 2015 10:39
yes, please first look at #108 + #110
on jruby-gradle-jar-plugin all my code is on the 'jruby-mains' branch and needs the above to PRs
Schalk W. Cronjé
@ysb33r
Apr 22 2015 10:43
OK let’s do the above then
Can I merge #108 then? It is pretty simple. Optimising it do do a single JVM execution in future can be done later.
Christian Meier
@mkristian
Apr 22 2015 10:46
yes please
Schalk W. Cronjé
@ysb33r
Apr 22 2015 10:47
OK, let’s see what travis and appveyor say
Christian Meier
@mkristian
Apr 22 2015 10:49
currently #108 fails with me locally but #110 runs through. FYI I am leaving to the airport in half an hour to go back to europe - so I will offline soon
Schalk W. Cronjé
@ysb33r
Apr 22 2015 10:50
sure. are you based in europe?
Christian Meier
@mkristian
Apr 22 2015 10:50
moved from england to germany 3 month ago
Schalk W. Cronjé
@ysb33r
Apr 22 2015 10:50
ah. Berlin?
Christian Meier
@mkristian
Apr 22 2015 10:51
but my son is half indian - so that is why he spends some of his vacations in india. no currently munich
you are in england if I understood it right ?!
Schalk W. Cronjé
@ysb33r
Apr 22 2015 10:52
cool. At least we’ll be in better TZs then. (yes, i’m inEngland, Aylesbury to be precise).
Christian Meier
@mkristian
Apr 22 2015 10:53
yes the difference between India and California is not so good :)
Schalk W. Cronjé
@ysb33r
Apr 22 2015 10:53
were you in India now?
Christian Meier
@mkristian
Apr 22 2015 10:54
bangalore
my wife sister life here
Schalk W. Cronjé
@ysb33r
Apr 22 2015 10:54
ah. I like Bangalore - been there many times.
Christian Meier
@mkristian
Apr 22 2015 10:54
ok
Schalk W. Cronjé
@ysb33r
Apr 22 2015 10:55
enjoy the 2h traffic to the airport :) (although you might just miss it, I guess).
Christian Meier
@mkristian
Apr 22 2015 10:56
yes, I was surprised that they want to go so early to airport - it is the traffic then
ok - the generate_dir_info fails with older jrubies.
either not generate them or pick a version for generating them which works. i.e. have fix jruby version for the internal tasks - and a configurable one for the exec
does this make sense ?
Schalk W. Cronjé
@ysb33r
Apr 22 2015 11:00
We’ll do two things I reckon. 1 Bump the suggested version in the release of base. 2 Put a check in to see if it is an odler version of ruby and skip the step, if it is.
Christian Meier
@mkristian
Apr 22 2015 11:00
ok
let me first try to catch that LoadError which happens when this method is missing.
Schalk W. Cronjé
@ysb33r
Apr 22 2015 11:03
Test probably fails because the integration test run with 1.7.11 (it is deliberate). I’ll loook into this.
Christian Meier
@mkristian
Apr 22 2015 11:05
yes this is understood. but the second execute could just rescue the load error on older jrubies. just running it locally.
is there a way to pick a specific integration test
instead of running all of them ?
Schalk W. Cronjé
@ysb33r
Apr 22 2015 11:06
-DintegrationTest.single=JRubyExecIntegrationSpec
Christian Meier
@mkristian
Apr 22 2015 11:06
yes - nice
Schalk W. Cronjé
@ysb33r
Apr 22 2015 11:06
gradle feature
Christian Meier
@mkristian
Apr 22 2015 11:07
still a newbie to gradle :)
and groovy
Schalk W. Cronjé
@ysb33r
Apr 22 2015 11:07
normally it would be -Dtest.single=, but beasue we are running tests out of the integrationTest configuration instead of test it has to be that. I’m sure you can recognise the pattern...
Christian Meier
@mkristian
Apr 22 2015 11:25
so I have push a fix for the jruby-1.7.11 problem and now I am off - see on friday I guess
Schalk W. Cronjé
@ysb33r
Apr 22 2015 11:28
Thnx. I’ll have a look.
R. Tyler Croy
@rtyler
Apr 22 2015 15:30
morning folks
Schalk W. Cronjé
@ysb33r
Apr 22 2015 15:30
Hello
I merged #108 this morning. It still fails a test, but for a good reason, not really a fault for the code. @mkristian says he did push a fix for that, but I have not seen it in the PR. It might still only be in his repo.
R. Tyler Croy
@rtyler
Apr 22 2015 15:36
@ysb33r you might want to look at his repo, he's on an international flight now from .in to .de
:P
Schalk W. Cronjé
@ysb33r
Apr 22 2015 15:42
I’ll look at it when I get home
in an hour or so
R. Tyler Croy
@rtyler
Apr 22 2015 15:43
I'll let you know if I get to it first, going through some gradle tasks this morning
word from gradleware is that they've fixed their bintray sync, so base 0.1.15 is 100% available now, not just through bintray
Schalk W. Cronjé
@ysb33r
Apr 22 2015 15:43
excellent
I wrote some code to guess the ruby version from the jarname and split it into the version triplet. There is a method there that just extracts the version string, so I’ve added another. This will help to fine tune incompatibilities within certain jruby versions. I’ll push that tonigh hopefully.
R. Tyler Croy
@rtyler
Apr 22 2015 15:46
for what issue? I'm not sure I understand
Schalk W. Cronjé
@ysb33r
Apr 22 2015 17:00
Remember when had to do that 1.7.14 workaround? That is one example. If the ,jrubydir stuff for instance does not work for ruby < 1.7.19 we would like to have an easy way to code that in
R. Tyler Croy
@rtyler
Apr 22 2015 17:06
vaguely
Schalk W. Cronjé
@ysb33r
Apr 22 2015 19:24
why o why o why, do I always stuggle with regexes in Java world
R. Tyler Croy
@rtyler
Apr 22 2015 19:31
hah, because they're awful
Schalk W. Cronjé
@ysb33r
Apr 22 2015 19:47
It’s that F’n Matcher object that always gets me. Even though Groovy simplifies it, it still gets me.
Schalk W. Cronjé
@ysb33r
Apr 22 2015 20:34
Right. I have a fix that will avoid creating the directory info if JRuby 1.7.18 or older is used. It will log a warning instead and then continue with the old behaviour
Schalk W. Cronjé
@ysb33r
Apr 22 2015 21:48
@rtyler I’ve created a PR to resolve that issue. I’m leaving it for you to review and you can merge it if you’re happy with my fix.
R. Tyler Croy
@rtyler
Apr 22 2015 22:03
:)
Schalk W. Cronjé
@ysb33r
Apr 22 2015 22:08
great. I can now have a look #110
Schalk W. Cronjé
@ysb33r
Apr 22 2015 22:31
@rtyler , do you know what @mkristian was trying to achieve with the maven layout in the JrubyPrepareJars task ?
R. Tyler Croy
@rtyler
Apr 22 2015 22:37
not entirely
does the jar bootstrap code use it
Schalk W. Cronjé
@ysb33r
Apr 22 2015 22:38
maybe. I wasn’t sure what that was doing yet - thought that by ensuring I understand what #110 does, I can would be able to undestand the jar changes later on
Schalk W. Cronjé
@ysb33r
Apr 22 2015 22:50
I’ll just add some more questions into the PR and leave it to him to answer on Friday.
R. Tyler Croy
@rtyler
Apr 22 2015 23:14
@ysb33r good code review notes, thanks for going through that
I can't wait for @mkristian to get back now :)