In bug 1231379 I disabled compression for omni.ja on nighlyt, which ends up reducing installer and update sizes. I just discovered that this doesn't work for l10n repacks because a different codepath is used to package omni.ja files there.

Review commit: https://reviewboard.mozilla.org/r/55482/diff/#index_header See other reviews: https://reviewboard.mozilla.org/r/55482/

try run is here: https://treeherder.mozilla.org/#/jobs?repo=try&revision=83d98b74ba2e&selectedJob=21516255 http://archive.mozilla.org/pub/firefox/try-builds/catlee@mozilla.com-83d98b74ba2e4389a72ad462fa9eb744ce802633/try-linux64/

Comment on attachment 8756956 [details] Bug 1275998 - L10n repacks should use the same compression settings for omni.ja as en-US https://reviewboard.mozilla.org/r/55482/#review52342 r+ with the following fixed. ::: python/mozbuild/mozpack/mozjar.py:408 (Diff revision 1) > + def is_compressed(self): > + ''' > + Return whether the jar archive is compressed. > + ''' > + return any(f.compressed for f in self) I'd rather not add this method, and just use any(f.compressed for f in jar) in UnpackFinder._open_jar. ::: python/mozbuild/mozpack/packager/l10n.py:240 (Diff revision 1) > l10n_finder = ComposedFinder(finders) > copier = FileCopier() > if app_finder.kind == 'flat': > formatter = FlatFormatter(copier) > elif app_finder.kind == 'jar': > formatter = JarFormatter(copier, optimize=app_finder.optimizedjars) You should add the compressed argument here too.

Comment on attachment 8756956 [details] Bug 1275998 - L10n repacks should use the same compression settings for omni.ja as en-US Review request updated; see interdiff: https://reviewboard.mozilla.org/r/55482/diff/1-2/

Pushed by catlee@mozilla.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/2c039c5822c2 L10n repacks should use the same compression settings for omni.ja as en-US r=glandium