I had to disable this test gardening today, since it was off by a pixel (and not flaking). I don't know if that was introduced in https://bugs.webkit.org/show_bug.cgi?id=95520 , or if this is something else, so I'm assigning it to levin (who originally wrote the test) and CCing jamesr (who landed 95520 yesterday). See https://bugs.webkit.org/show_bug.cgi?id=95571 for the patch disabling this test, and http://build.chromium.org/p/chromium.webkit/builders/Webkit%20Win/builds/17247 for a fine sample of a failed build.
Gavin, would you add the regression range (on the WebKit and Chromium side) to aid anyone who will look at this. Unassigning from me because it won't get fixed if you assign it to me since my time on WebKit/Chromium has been nearly non-existant for months. If there is a breakage, it is mostly likely going to affect notifications or the autosized popup from the browser icons that extensions pop-up, so I'm cc'ing folks who may be most concerned about that. I'm more than happy to give help/pointers/etc. to anyone else taking this one. PS I can't image how https://bugs.webkit.org/show_bug.cgi?id=95520 could have broken this.
Thanks David. I can't imagine either, I just couldn't see a better candidate. In webkit, the range is r127171 to 127194. In chromium, the range is r154235 to r154278.
95520 is a refactor of code that's not used yet, so I doubt it's related.
Test related bugs being marked WontFix. TestExpectations or a disabled comment will contain bug# if still relevant.