Bugzilla

(In reply to Botond Ballo [:botond] from comment #7)
> (In reply to Glenn Watson [:gw] from comment #4)
> > Alternatively I could land my patch with the pref disabled, and change the test expectations, and the pref could be enabled at a later time when APZ team can investigate those perma-fails?
> 
> I would suggest going with this approach.
> 
> I assume by "change test expectations" you mean adding `pref(layout.scroll.disable-pixel-alignment,true)` to the `snapped-scrolled-content-1.html` test cases, to validate going forward that they are passing in this configuration even though this is not yet the default configuration.

Yes, that's right. I've updated the patch if you can review the test changes before I land. To clarify:

* Added `pref(layout.scroll.disable-pixel-alignment,true)` to all of that block of tests.
* Added `fails-if(useDrawSnapshot)` to 3 tests - these tests fail in `drawSnapshot` mode with the preference enabled and _without_ the WR change present.
* Added `fails-if(useDrawSnapshot)` to 2 tests - these tests were unconditional FAILs, but not pass in WR mode with the patch (continue to fail in `drawSnapshot` mode).

(In reply to Botond Ballo [:botond] from comment #7)
> (In reply to Glenn Watson [:gw] from comment #4)
> > Alternatively I could land my patch with the pref disabled, and change the test expectations, and the pref could be enabled at a later time when APZ team can investigate those perma-fails?
> 
> I would suggest going with this approach.
> 
> I assume by "change test expectations" you mean adding `pref(layout.scroll.disable-pixel-alignment,true)` to the `snapped-scrolled-content-1.html` test cases, to validate going forward that they are passing in this configuration even though this is not yet the default configuration.

Yes, that's right. I've updated the patch if you can review the test changes before I land. To clarify:

* Added `pref(layout.scroll.disable-pixel-alignment,true)` to all of that block of tests.
* Added `fails-if(useDrawSnapshot)` to 3 tests - these tests fail in `drawSnapshot` mode with the preference enabled and _without_ the WR change present.
* Added `fails-if(useDrawSnapshot)` to 2 tests - these tests were unconditional FAILs, but now pass in WR mode with the patch (continue to fail in `drawSnapshot` mode).