These are chat archives for rgladwell/m2e-android

29th
Jun 2015
Csaba Kozák
@WonderCsabo
Jun 29 2015 10:15
@fbricon when do you plan to release 1.6.1?
Ricardo Gladwell
@rgladwell
Jun 29 2015 10:17
Is there a p2 repo for 1.6.1?
Csaba Kozák
@WonderCsabo
Jun 29 2015 10:18
there is, several comments earlier the milestone repo, i already tested that
Ricardo Gladwell
@rgladwell
Jun 29 2015 10:18
Sweet
Csaba Kozák
@WonderCsabo
Jun 29 2015 10:19
should we use that, or wait until 1.6.1 is not released?
Ricardo Gladwell
@rgladwell
Jun 29 2015 10:22
I thin to get the build green we should use the temp p2 repo
Csaba Kozák
@WonderCsabo
Jun 29 2015 10:23
okay
i am finishing my runsAfter branch, and push a PR
Ricardo Gladwell
@rgladwell
Jun 29 2015 10:25
Cool thanks for that
Csaba Kozák
@WonderCsabo
Jun 29 2015 10:25
well, on the second thought that will not fix the build :(
we have to implement that classpath listener thing
Ricardo Gladwell
@rgladwell
Jun 29 2015 10:26
I guess to get the build to go green
Csaba Kozák
@WonderCsabo
Jun 29 2015 10:27
i also mean that
the build is failing currently because "Android container should not be exported"
that can be only fixed by the classpath listener
#360
Ricardo Gladwell
@rgladwell
Jun 29 2015 10:37
Will #360 appear to make the build go green, but it wont be fixed until the other issues will be resolved?
Csaba Kozák
@WonderCsabo
Jun 29 2015 10:38
#360 will fix that issue itself
but m2e 1.6 support is not complete without #359
Ricardo Gladwell
@rgladwell
Jun 29 2015 10:40
Coolio
Csaba Kozák
@WonderCsabo
Jun 29 2015 10:41
i would not say that, we have to create an ugly implementation in #360 instead of the current nice one which will also take time
Ricardo Gladwell
@rgladwell
Jun 29 2015 10:44
I sometimes wonder if we're the only people supporting a third party m2e extension
Csaba Kozák
@WonderCsabo
Jun 29 2015 10:44
not at all, but please note that #360 is needed because of the ADT plugin, not m2e
and i would not blame them, since they are just creating a classpath container in the standard why
Ricardo Gladwell
@rgladwell
Jun 29 2015 10:46
Well m2e did make major changes to the way their plugin worked, they are not free of blame
Csaba Kozák
@WonderCsabo
Jun 29 2015 10:46
yeah, they should indicate that to m2e connector developers
but regarding the ADT container: we were just lucky, that the m2e configuration cycle was longer then the container init...
Ricardo Gladwell
@rgladwell
Jun 29 2015 10:52
This project has always been a collection of hacks unfortunately, due to the problems with our upstream projects
Csaba Kozák
@WonderCsabo
Jun 29 2015 10:52
that is true
Csaba Kozák
@WonderCsabo
Jun 29 2015 13:01
@rgladwell i am starting to think that we should not bother with Android Private Libraries container
that container contains jars which are added to the libs folder in the project, but such folder should not be in Maven projects and users should utilize Maven dependency management. Actually this container is always empty in our test projects.
Ricardo Gladwell
@rgladwell
Jun 29 2015 13:07
I’m not keen to introduce a regression like that, some of our users maybe depending on this feature.
Csaba Kozák
@WonderCsabo
Jun 29 2015 13:08
but in that case the container actually should be exported, shouldn't?
Ricardo Gladwell
@rgladwell
Jun 29 2015 13:10
i think it was a use-case where a user needed both types of dependencies, possible because he was sharing projects with Android Studio
So in this case the Android container lib should never be exported if used with a Maven-enabled IDE
Csaba Kozák
@WonderCsabo
Jun 29 2015 13:11
i cannot follow your reasoning, @rgladwell
Ricardo Gladwell
@rgladwell
Jun 29 2015 13:11
which bit don’t you follow?
Csaba Kozák
@WonderCsabo
Jun 29 2015 13:12
you mean the same jars were both present in the container and the maven dependencies container?
in case of the user you are talking about
Ricardo Gladwell
@rgladwell
Jun 29 2015 13:13
They were using Maven for dependency management and the libs/ folder i guess
it was a long time ago
same depedencies in both maybe?
Csaba Kozák
@WonderCsabo
Jun 29 2015 13:14
also please note you introduced this configurator in #155, but you never really documented it why
Ricardo Gladwell
@rgladwell
Jun 29 2015 13:15
I think I moved it as part of that issue, I think I created it in an earlier commit.
Csaba Kozák
@WonderCsabo
Jun 29 2015 13:15
@rgladwell i see, but we should not support such a scenario, since Gradle has excellent dependency management, also the user could just add the jars to another folder than libs
Ricardo Gladwell
@rgladwell
Jun 29 2015 13:16
As I said before, I’m not sure if that’s true and I’m concerned we may break this issue for some of our users because we don’t know exactly how our users are using the plugin.
Csaba Kozák
@WonderCsabo
Jun 29 2015 13:19
@rgladwell honestly i think that is a really bad scenario, and i see no reason to support it. Moreover, i tried to implement the classpath listener based solution, but it appears to be very hard.
Ricardo Gladwell
@rgladwell
Jun 29 2015 13:22
I’m not sure that is the scenario. I think I had to do this as part of a larger change to support a major feature. The classpath container in question was created by me as part of a patch to the ADT.
But there are other reasons to support it: not breaking features for our users is one reason.
Csaba Kozák
@WonderCsabo
Jun 29 2015 13:23
You mean the Android Private Libraries container was contributed by you?
Ricardo Gladwell
@rgladwell
Jun 29 2015 13:23
I think so, it was a long time ago
It took 8 monthes to get the ADT DEV team to merge it
Csaba Kozák
@WonderCsabo
Jun 29 2015 14:19
@rgladwell i think i have a fix
Fred Bricon
@fbricon
Jun 29 2015 14:38
have you guys tested the staged m2e 1.6.1?
Csaba Kozák
@WonderCsabo
Jun 29 2015 14:38
yeah, it works well
Fred Bricon
@fbricon
Jun 29 2015 14:39
good, I'll release it later today
Csaba Kozák
@WonderCsabo
Jun 29 2015 14:45
great!