Issue 4697 - Add support for WebExtension-based Firefox development builds

Issue 4697 - Add support for WebExtension-based Firefox development builds

[Wladimir Palant, 2016-12-05 08:42:03]

https://codereview.adblockplus.org/29366797/diff/29366798/sitescripts/extensi...
File sitescripts/extensions/bin/createNightlies.py (right):

https://codereview.adblockplus.org/29366797/diff/29366798/sitescripts/extensi...
sitescripts/extensions/bin/createNightlies.py:250: templateName =
'geckoUpdateManifest'
Note: the builds currently won't use that update manifest but we'll take care of
it.

https://codereview.adblockplus.org/29366797/diff/29366798/sitescripts/extensi...
sitescripts/extensions/bin/createNightlies.py:571:
get_config().has_option('extensions', 'amo_key')):
Overindenting seems to be the most readable approach that doesn't cause flake8
to complain.

[Wladimir Palant, 2016-12-05 09:49:19]

I realized that the update manifest format changed, so Patch Set 2 now
implements the format as documented under
https://developer.mozilla.org/en-US/Add-ons/Updates.

https://codereview.adblockplus.org/29366797/diff/29366798/.sitescripts.example
File .sitescripts.example (right):

https://codereview.adblockplus.org/29366797/diff/29366798/.sitescripts.exampl...
.sitescripts.example:106: geckoUpdateManifestPath=%(root)s/www/update.rdf
This configuration setting is no longer used.

[Jon Sonesen, 2016-12-05 10:45:54]

On 2016/12/05 09:49:19, Wladimir Palant wrote:
> I realized that the update manifest format changed, so Patch Set 2 now
> implements the format as documented under
> https://developer.mozilla.org/en-US/Add-ons/Updates.
> 
> https://codereview.adblockplus.org/29366797/diff/29366798/.sitescripts.example
> File .sitescripts.example (right):
> 
>
https://codereview.adblockplus.org/29366797/diff/29366798/.sitescripts.exampl...
> .sitescripts.example:106: geckoUpdateManifestPath=%(root)s/www/update.rdf
> This configuration setting is no longer used.

Just quickly testing these changes it looks like the removal of the update.rdf
file causes the test_UpdateManifests.py tests to fail. This can be fixed by
replacing the rdf2data method in that script with a method that reads the file
into a dict.

[Wladimir Palant, 2016-12-05 11:22:08]

On 2016/12/05 10:45:54, Jon Sonesen wrote:
> Just quickly testing these changes it looks like the removal of the update.rdf
> file causes the test_UpdateManifests.py tests to fail.

Ouch, I meant to update the tests but somehow forgot. Fixed those now, and also
updateUpdateManifests script even though I suspect that Gecko manifest
generation for release builds is dead code.

[Jon Sonesen, 2016-12-05 12:02:23]

On 2016/12/05 11:22:08, Wladimir Palant wrote:
> On 2016/12/05 10:45:54, Jon Sonesen wrote:
> > Just quickly testing these changes it looks like the removal of the
update.rdf
> > file causes the test_UpdateManifests.py tests to fail.
> 
> Ouch, I meant to update the tests but somehow forgot. Fixed those now, and
also
> updateUpdateManifests script even though I suspect that Gecko manifest
> generation for release builds is dead code.

Ok, thanks for getting that switched around. Perhaps I can do some digging to
see if the code is in fact dead. If so it would be good to comment on that in
the code to avoid maintaining code that is not in use. But this discussion is
probably out of scope for codereview.  LGTM.