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

26th
Aug 2014
R. Tyler Croy
@rtyler
Aug 26 2014 00:22
@ysb33r I'm also setting up an organization on bintray properly so we can both release changes moving forward
I think we'll likely end up working on different sections of the plugins
R. Tyler Croy
@rtyler
Aug 26 2014 05:18
alright I think that's enough house-keeping to get us started with @jruby-gradle for tonight, should be back real soon! :)
@ysb33r I wanted to wait for you to chime in on #44 before I renamed a boatload of files, just to sanity check things
moving to the jruby-gradle org on @bintray also means we can reset the version counter to 0.1.0 until we get the code moved into jruby-gradle-{j,w}ar-plugins
rtyler @rtyler falls asleep
R. Tyler Croy
@rtyler
Aug 26 2014 13:47
@ysb33r morning :)
Schalk W. Cronjé
@ysb33r
Aug 26 2014 14:01
hello
R. Tyler Croy
@rtyler
Aug 26 2014 14:08
just got warbler-bootstrap into jcenter from the jruby-gradle org on bintray
Schalk W. Cronjé
@ysb33r
Aug 26 2014 14:08
One less repo to load in JRubyPlugin :)
R. Tyler Croy
@rtyler
Aug 26 2014 14:09
\o/
I figure the libraries repo can be for supporting tooling like warbler-bootstrap
then the plugins repo for the actual plugins themselves
Schalk W. Cronjé
@ysb33r
Aug 26 2014 14:10
Indeed
R. Tyler Croy
@rtyler
Aug 26 2014 14:12
do you have changes in flight, or could i safely resolve #44 and reset the version counter?
Schalk W. Cronjé
@ysb33r
Aug 26 2014 14:13
Nothing new

45 was the only thing

R. Tyler Croy
@rtyler
Aug 26 2014 14:13
heh
silly markdown
Schalk W. Cronjé
@ysb33r
Aug 26 2014 14:13
indeed
R. Tyler Croy
@rtyler
Aug 26 2014 14:13
okay, I'm going to make those changes on the bus
I'll be back online in a few
rtyler @rtyler is really liking the separate plugin pattern
Schalk W. Cronjé
@ysb33r
Aug 26 2014 14:14
I’ll let you get on with it. I have no new fixes planned
Are you going with multiple repositories instead of multi-level gradle build?
R. Tyler Croy
@rtyler
Aug 26 2014 14:34
multiple repos yeah
I really think that the jruby-gradle-plugin by itself will move much much slower than the others
I've also seen the gradle-plugins repository (IIRC) and I personally think it's very disorganized/hard to navigate
Schalk W. Cronjé
@ysb33r
Aug 26 2014 14:48
well, if you look at the gradle source code - that does feel over the top. Smaller projects tend to work well with multi-legel builds. But if you want to use multi-repos that’s fine as well. We’ll have to duplicate some stuff in the gradle build scripts, but probably the worse of it
and if the base plugin grows slow, it might be better to run it as seperate repos
rtyler @rtyler nods
R. Tyler Croy
@rtyler
Aug 26 2014 14:49
I already played the copy/paste between build.gradle game
but of course, we can always fix that with an internal gradle plugin :P
can you not use dashes in a package name?
Schalk W. Cronjé
@ysb33r
Aug 26 2014 14:50
I’ve done both as well. When I started off with some gradle plugins, I put them in one repo, but then found that I had to split some of them off into other repos later on
"can you not use dashes in a package name” Where? in gradle or bin tray?
R. Tyler Croy
@rtyler
Aug 26 2014 14:51
gradle
or java
looks like no
Schalk W. Cronjé
@ysb33r
Aug 26 2014 14:52
are referring to packagename such as jruby-gradle-1.2.3.jar?
or something else
R. Tyler Croy
@rtyler
Aug 26 2014 15:15
@ysb33r org.github.jruby-gradle
Schalk W. Cronjé
@ysb33r
Aug 26 2014 15:16
ah
nope, it should not work because that reflects a java namespace which cannot contain ‘-'
rtyler @rtyler nods
R. Tyler Croy
@rtyler
Aug 26 2014 15:17
what should the package name be then?
Schalk W. Cronjé
@ysb33r
Aug 26 2014 15:17
although I suspect it is a Java limitation not a JVM bytecode one
R. Tyler Croy
@rtyler
Aug 26 2014 15:34
ugh, this renaming of things is a pain in the ass
R. Tyler Croy
@rtyler
Aug 26 2014 15:43
@ysb33r for the base plugin, should we just call it base?
com.github.jruby-gradle.jruby-gradle-plugin feels too verbose
I kind of like com.github.jruby-gradle.base and com.github.jruby-gradle.war etc
Schalk W. Cronjé
@ysb33r
Aug 26 2014 15:50
there is a difference between the java namespace and the gradle plugin namespace
R. Tyler Croy
@rtyler
Aug 26 2014 15:50
I know
Schalk W. Cronjé
@ysb33r
Aug 26 2014 15:50
I would have thought to have left the java namespace alone being com.lookout.jruby.*
R. Tyler Croy
@rtyler
Aug 26 2014 15:50
eh, I wanted to rip that out
my employer isn't really part of this project
I just didn't want to use a personal domain :P
Schalk W. Cronjé
@ysb33r
Aug 26 2014 15:51
OK then
com.github.jruby.gradle as the top-layer
as by example most of my other plugins are in org.ysb33r.gradle namespace
R. Tyler Croy
@rtyler
Aug 26 2014 15:52
but we're not part of the jruby org? is com.github.jruby-gradle.base that bad? :(
I also thought putting gradle there wasn't something you were supposed to do
rtyler @rtyler points to http://plugins.gradle.org/submit
Schalk W. Cronjé
@ysb33r
Aug 26 2014 15:53
right, I was getting to that
For the Java namespace I would have suggested com.github.jruby.gradle, BUT as you pointed out there, might be an issue with it. For the gradle plugin portal I would go with com.github.jruby-gradle.(base|war) etc.
the problem with the Java namespace is that the dash is not allowed as you have discovered
R. Tyler Croy
@rtyler
Aug 26 2014 15:55
java namespace in my changes right now is com.github.jrubygradle
I think this will work, changing tests now
Schalk W. Cronjé
@ysb33r
Aug 26 2014 15:55
I can live with that
As to my earlier example - I used org.ysb33r.gradle as the java toplevel for most of my plugins, but on the portal I just drop the gradle part so they become something like org.ysb33r.doxygen.
rtyler @rtyler nods
R. Tyler Croy
@rtyler
Aug 26 2014 16:00
running integration tests for the first time in a long time :P
Schalk W. Cronjé
@ysb33r
Aug 26 2014 16:01
5mins on my laptop
after making that GemUtils change that you merged today
rtyler @rtyler nods
R. Tyler Croy
@rtyler
Aug 26 2014 16:02
my laptop is a bit older, ~4 years :P
R. Tyler Croy
@rtyler
Aug 26 2014 16:09
~8 minutes
not shabby
R. Tyler Croy
@rtyler
Aug 26 2014 16:15
@ysb33r @buildhive will look at #46, but I'd love a review from you before merging
Schalk W. Cronjé
@ysb33r
Aug 26 2014 16:21
will do
Schalk W. Cronjé
@ysb33r
Aug 26 2014 16:28
looks good enough to me. Did you do all the renaming with vi?
R. Tyler Croy
@rtyler
Aug 26 2014 16:28
heh, yes
i'm not cool enough for an IDE
Schalk W. Cronjé
@ysb33r
Aug 26 2014 16:29
:)
I use a variety of editors, but I would have done that renaming with IntelliJ
R. Tyler Croy
@rtyler
Aug 26 2014 16:30
@ysb33r I'm going to release this as is to @bintray so I can get the linking to jcenter going, while we wait for that, I'll pull the war pieces into the war plugin
Schalk W. Cronjé
@ysb33r
Aug 26 2014 16:36
Sure.
I am building your first merge right now as well
R. Tyler Croy
@rtyler
Aug 26 2014 16:38
alright, tagged and released properly
requested inclusion in jcenter too
Schalk W. Cronjé
@ysb33r
Aug 26 2014 16:50
I actually think you should bump the release version to 3.0 - major version signifies major restructuring and breaking backwards compatiblity
That is when the first real release comes out
quite a bit of documentation to be rewritten, especially the README
R. Tyler Croy
@rtyler
Aug 26 2014 16:54
the plugin is completely differently namespaced though
there's never been a release other than 0.1.0 of com.github.jruby-gradle.base before
Schalk W. Cronjé
@ysb33r
Aug 26 2014 16:55
Mmmm… trying to think in terms of the user here, but maybe a 1.0 release would work
R. Tyler Croy
@rtyler
Aug 26 2014 16:55
that's why I reset the version counter
I wanted to create 1.0 once the jar and war code comes out
Schalk W. Cronjé
@ysb33r
Aug 26 2014 16:56
I think it is far enough beyond a 0.x release
Agreed
R. Tyler Croy
@rtyler
Aug 26 2014 16:56
I didn't want to create a fast-follow 2.x when I gut that, which should be sometime today if I'm lucky
Schalk W. Cronjé
@ysb33r
Aug 26 2014 16:57
When one 1.0 is out, I’ll submit a pull request to asciidoctorj for a conversion to gradle
R. Tyler Croy
@rtyler
Aug 26 2014 17:00
woooot
@ysb33r asciidoctor doesn't depend on jar or war tasks does it?
Schalk W. Cronjé
@ysb33r
Aug 26 2014 17:06
nope.
Will go without it. See https://t.co/Ad8riIzEOg
R. Tyler Croy
@rtyler
Aug 26 2014 17:13
clever ivy hack
R. Tyler Croy
@rtyler
Aug 26 2014 17:27
wonder how long it's going to take for this to get accepted into jcenter
I usually go to sleep and wake up to binaries accepted :P
Schalk W. Cronjé
@ysb33r
Aug 26 2014 17:28
It’s about right, given the support usually comes out of Tel Aviv.
R. Tyler Croy
@rtyler
Aug 26 2014 18:14
WOOHOOO included in jcenter
R. Tyler Croy
@rtyler
Aug 26 2014 18:57
not showing up on plugins.gradle.org
wonder if I messed something up
Schalk W. Cronjé
@ysb33r
Aug 26 2014 19:08
the plugin is not yet linked to the Gradle Bintry repo
R. Tyler Croy
@rtyler
Aug 26 2014 19:57
OH
right
alright sent that request too
Schalk W. Cronjé
@ysb33r
Aug 26 2014 20:22
Besides, GEM_HOME, Is there any way of telling jruby the location of extracted gem directories?
R. Tyler Croy
@rtyler
Aug 26 2014 20:45
there's a RUBY_LIB environment variable IIRC
R. Tyler Croy
@rtyler
Aug 26 2014 21:07
this stupid gradleware bot is taking forever to approve this sucker
R. Tyler Croy
@rtyler
Aug 26 2014 22:19
plugins { id "com.github.jruby-gradle.base" version "0.1.0" }
so cooooooool
rtyler @rtyler does a happy dance
Schalk W. Cronjé
@ysb33r
Aug 26 2014 22:44
I need to get some sleep.
R. Tyler Croy
@rtyler
Aug 26 2014 22:47
nighty night
Schalk W. Cronjé
@ysb33r
Aug 26 2014 22:52
go now, but got one more question: I noticed when running jruby one can pass a -I parameter. It refers to the load path. How does that differ from GEM_HOME?
R. Tyler Croy
@rtyler
Aug 26 2014 22:52
LOAD_PATH is something that applies to the require machinery in Ruby
GEM_HOME is used by rubygems, which is related to require machinery but not part of core Ruby
both can be used to make a require 'sinatra' work
Schalk W. Cronjé
@ysb33r
Aug 26 2014 22:55
I’m quite confused on those and cannot seem to find a good reference that explains it.
In both Perl & Python one can link search paths together to look for modules using PERL5LIB or PYTHONPATH or one can supply -I on the command-line. I keep thinking GEM_HOME should work the same. Or is LOAD_PATH more like this?
R. Tyler Croy
@rtyler
Aug 26 2014 22:57
PYTHONPATH is like LOAD_PATH
GEM_HOME is more like whatever setuptools uses for it's wizardry
Schalk W. Cronjé
@ysb33r
Aug 26 2014 22:58
Maybe the tests should place LOAD_PATH in the test environment rather than GEM_HOME.
R. Tyler Croy
@rtyler
Aug 26 2014 23:10
that would work