Create skia_webkit.gyp to improve rebaselining.
Created attachment 169479 [details] Patch
If this is committed, then https://codereview.chromium.org/11191074 would be checked in to Chromium after the next WebKit roll into Chromium.
Am I correct in understanding that there is a matching chromium patch that adds the reference to this file? To rebaseline, we would: 1. Change a flag in this new skia_webkit.gyp file 2. Watch the build bots generate all the new images (with lots and lots of failures) 3. Rebaseline WebKit 4. Roll WebKit into Chromium, and pick up both the flag change and the new baselines Sounds to me like a much better workflow. The only downside is that it adds the possibility that a Skia change in WebKit might cause breakage of chromium browser or unit tests on the WebKit roll. I don't think we can really do anything much about that, but it does mean that the skia team will have to be aware when a flag change rolls into chromium.
>Am I correct in understanding that there is a matching chromium patch that adds the reference to this file? See https://codereview.chromium.org/11191074 >To rebaseline, we would: Your description is accurate as to the intended workflow. The entire process previous to this is something like. -4. Make a change in Skia behind a build flag because it is known to need new baselines in WebKit. -3. Roll Skia into Chromium with the code suppression flag in Chromium's skia.gyp. -2. At any point here or previous, get the code suppression flag into the skia_webkit.gyp and roll (or wait for roll of) WebKit into Chromium. -1. Remove suppression from Chromium's skia.gyp. 0. Roll (or wait for roll of) Chromium into WebKit. At this point the rebaseline can proceed as you describe. >might cause breakage of chromium browser or unit tests on the WebKit roll To mitigate this, we can send the full change to the trybots ahead of time to make sure that only layout tests are affected.
This change (and the corresponding Chromium-side change) look fine (well, frankly, it looks ugly but I don't think there's another way to solve this), but I'd like to hold off on landing it until after we can meet to discuss the flow on Wednesday, if that's okay.
Comment on attachment 169479 [details] Patch We agreed to land it :).
Comment on attachment 169479 [details] Patch Clearing flags on attachment: 169479 Committed r132385: <http://trac.webkit.org/changeset/132385>
All reviewed patches have been landed. Closing bug.