Currently, the caret in cursor mode is always hide after scrolling. I
think it's better to restore its appearance if it's logically visible
before scrolling. That is:
1) If the caret is logically visible (Normal or NormalNotShown) before
scrolling, update carets on scroll-end to show the caret again.
2) If the caret is hide due to timeout or blur, it should remain hidden
on scroll-end.
--HG--
extra : transplant_source : %40%2C%95%CB%EDzt%FB%F0CpK%C94TN2%27%2A%A3
(This patch renames the function to "ParseWebkitPrefixedGradientWithService", indicating that this is the version that calls out to the CSSUnprefixingService.)
The conversion is as follows:
- GraphicsFilter::FILTER_NEAREST == gfx::Filter::POINT
- GraphicsFilter::FILTER_GOOD == gfx::Filter::GOOD
- GraphicsFilter::FILTER_BEST == gfx::Filter::LINEAR
Also typedef GraphicsFilter to gfx::Filter; this will be removed in the next
patch.
These changes mean ToFilter() and ThebesFilter() are no longer needed.
TouchCaret does not have this issue since it clamps the dragging point
to the editable content boundary.
Fix this bug by porting TouchCaret::GetContentBoundary() to
AccessibleCaret. I apply the clamp logic to both cursor mode and
selection mode if the focus node is on an editable content, which makes
carets dragging in selection mode smoother than SelectionCarets.
--HG--
extra : commitid : CbqEsabADIl
extra : rebase_source : 86847ced02859cd556806d547461e8f4167eab73
Test the second carets can still be dragging after its appearance
changing from Normal to NormalNotShown then back to Normal again. This
test is only for AccessibleCaret, not for SelectionCarets.
--HG--
extra : commitid : CbqEsabADIl
extra : rebase_source : 7bdf94164c25209c92fc4bd5fa4ae9e83a498bf7
word_location() did not work if there are multiple spaces between words.
We split by \S+ which is non-spaces, so tokens[0] is an space token.
Test cases are added to ensure the correctness.
--HG--
extra : commitid : CbqEsabADIl
extra : rebase_source : 9afccc28c8289244871b3912435a1ee63c3b4c5d
To get the same behavior across all platforms for carets test, disable
'layout.word_select.eat_space_to_next_word'. In this way, we don't need
to worry about the spaces being selected on Windows, and those strip()
in individual tests can be eliminated.
--HG--
extra : commitid : CbqEsabADIl
extra : rebase_source : df60d7092ecf95baef64fff4706dd189fbb11b9e
This patch renames AnimationCollection::mNeedsRefreshes to indicate that it
no longer has any relationship to whether or not we observe the refresh driver.
We need to do this so effects can query their owning animation for the current
time and avoid falling out of sync. Furthermore, this pointer is needed
for a number of other bugs (e.g. bug 1166500 comment 12, or bug 1190235)
anyway.
Note that this does not emulate the WebKit quirk of supporting
element.style["-webkit-animation"] or supporting the uppercase
element.style.WebkitAnimation (etc.) as opposed to the lowercase (and
enumerable) element.style.webkitAnimation.
Note that this also does not add aliases for transition or animation
events.
(Patch has been minorly tweaked by dholbert, with dbaron's approval.)