As debug builds are inherently slow, we need to increase the wait
time before triggering the `setTimeout` callback so that the next
`marionette.execute_script` statement has time to run.
MozReview-Commit-ID: IAgicNAmVA6
--HG--
extra : rebase_source : dcb0ae95265e28e08c257260eff426d87a6340af
Add a target property and show the event name (and modifiers if any) in the title.
Minor edit in GripRep to allow to pass a title as a prop.
Edit tests accordingly, and add some to test node highlighting.
MozReview-Commit-ID: JmMidejKS6m
--HG--
extra : rebase_source : 29ad4d398ad7518d752b323a6e6393cd30b18a0a
Ideally we would have used the GUID to join fields to build the compound key but because we show the uniqueId in the tree that would lead to an ugly user experience. As an
alternative we simply use e.g. db1 (default) as the uniqueId. This works as a unique key, is informative for the user and means minimal changes to tests.
MozReview-Commit-ID: I0XxTIoV33p
--HG--
extra : rebase_source : a8a9729ae639f68e849fc8239c36f451496847ad
The class ViewportFrame doesn't have ns-prefix, so it's better to drop the
ns-prefix in the file names to avoid confusion.
MozReview-Commit-ID: 8Jrmfzb3tVR
--HG--
rename : layout/generic/nsViewportFrame.cpp => layout/generic/ViewportFrame.cpp
rename : layout/generic/nsViewportFrame.h => layout/generic/ViewportFrame.h
extra : rebase_source : d2c6a80275d696a2886983a4bbb46821784c7fe6
Need to add #includes "nsPresContext.h" and "nsThreadUtils.h" in
PresShell.h, and "nsViewportInfo.h" in MobileViewportManager.h to compile.
MozReview-Commit-ID: F8bmOwllOiH
--HG--
extra : rebase_source : cacbb30d854c68ea720f849d03b649bbd3c5da03
The class PresShell doesn't have ns-prefix, so it's better to drop ns-prefix
in the file names to avoid confusion.
MozReview-Commit-ID: IljxsF5CVjh
--HG--
rename : layout/base/nsPresShell.cpp => layout/base/PresShell.cpp
rename : layout/base/nsPresShell.h => layout/base/PresShell.h
extra : rebase_source : e73a6873f360a7223d87bd22d9854fc9603c499f
Update the tooltool manifests for the android builds to include
support for the i686-linux-android target.
MozReview-Commit-ID: EyALhnfG4Kz
--HG--
extra : rebase_source : a85b8c1509458e1f5a8f8eae163e38edd1c363ce
Previously for the spacing at the top and bottom of the tabs grid panel we had
the total desired vertical spacing set on the RecyclerView padding, but then
there was also an additional half spacing coming from the ItemDecoration padding
in the top and bottom rows. Here we decrease the RecyclerView vertical padding
to account for the ItemDecoration vertical padding.
The alternative route of keeping full RecyclerView padding and then having the
ItemDecoration adjust itself depending on its position turned out to not "just
work": for example, if span count is three and you have four tabs, then the
first three tabs have half padding along their bottoms (since there are two
rows), but then when you close tab 4, there's now only one row and so the three
tabs should have no padding along their bottoms (since the RecyclerView already
has its own full padding), *but only tab 3 gets its ItemDecoration updated
automatically*, so it gets 0 bottom padding, but its row still has other tabs
with half bottom padding, so a) there's still too much bottom padding in that
row coming from tabs 1 and 2, and b) tab 3 sits too far down in its space
because it doesn't have the bottom padding that tabs 1 and 2 do.
That issue could be fixed by updating all ItemDecorations after each close, but
the patch here is both simpler and leads to less runtime work.
MozReview-Commit-ID: 2WeZ6QdfIF4
--HG--
extra : rebase_source : 85aec8adfdaacac7062827e273cf697c09167d63
When we jump to fullscreen, the OS sends a WM_MOUSELEAVE, which we turn into a eMouseExitFromWidget. The eMouseEnterIntoWidget was previously only sent when the mouse moved into the widget, which required the mouse to move. When entering fullscreen, we want the eMouseEnterIntoWidget to happen right away.