cordova-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Michal Mocny <mmo...@chromium.org>
Subject Re: Medic/Buildbot - building release versions
Date Thu, 17 Oct 2013 14:31:25 GMT
I think using MS at master with released platform/cli was a powder keg
waiting to explode.

We shipped a broken 3.1 mobile spec, and so +1 to your suggestion of fixing
it and moving CI back to test testing cadence releases with the matching
MobileSpec.

Side note: part of the problem seems to be that mobile spec is both an app
and a set of plugins, and those plugins (namely the dependency plugin)
weren't being updated on the plugin update schedule (weekly).  We should
make a note to do so in the future (or even move MS plugins out to another
repo and add to registry?)

-Michal


On Thu, Oct 17, 2013 at 8:40 AM, David Kemp <drkemp@chromium.org> wrote:

> TL;DR : Android release 3.1 is hard to test again due to added tests (to
> master) that fail. patching 3.1 mobile-spec would be a good thing.
>
> Our plan here was to continuously build iOS and Android, both current
> release and master (4 builds). To do this we build:
> IOS_Master : iOS (master), plugins(dev), mobile-spec(master)
> IOS_Release : iOS (3.1), plugins(master), mobile-spec(master)
> Android_Master : iOS (master), plugins(dev), mobile-spec(master)
> Android_Release : iOS (3.1), plugins(master, mobile-spec(master)
>
> The reason that mobile-spec master is used for release branches is because
> the plugins were renamed just before 3.1, but the dependencies listed in
> mobile-spec were not updated.
> (cordova-mobile-spec/dependencies-plugin/plugin.xml)
>
> To avoid that problem, we switched to testing 3.1 with master mobile-spec.
> Until last night that worked.
>
> However, now mobile-spec (master) has new tests added that will always fail
> for 3.1.
>
> SO, I would like to patch 3.1 mobilespec with the plugin dependencies from
> master.
>
> Any concerns?
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message