webkit-patch rebaseline-expectations should take a --platform arg
Created attachment 165711 [details] Patch
Note that this still needs a test ...
Created attachment 172195 [details] Patch
unfortunately, because of the way the code is structured (with the parallel fork of rebaseline-test-internal) it's hard to really test this very well. I don't think the unit tests for --platform=* add much. I did test a bunch by hand (which is how I discovered the locking problem). Let me know if you want me to split the locking code out into a separate patch.
Comment on attachment 172195 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=172195&action=review > Tools/Scripts/webkitpy/tool/commands/rebaseline.py:157 > + try: Maybe add a FIXME to handle these the same way we do the SCM updates?
(In reply to comment #5) > (From update of attachment 172195 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=172195&action=review > > > Tools/Scripts/webkitpy/tool/commands/rebaseline.py:157 > > + try: > > Maybe add a FIXME to handle these the same way we do the SCM updates? Will do. I'm actually thinking about reworking the way we do parallelism differently; given that the slow part is fetching the URLs, maybe we should just do that in parallel and do the rest of the work all at once (as if we were rebaselining locally). I'll probably fool around with this approach and see how it goes ...
Committed r133508: <http://trac.webkit.org/changeset/133508>