The accessible caret manager is owned by the event hub, that is owned by the
shell.
All the callers of methods that call FlushLayout on the AccessibleCaretManager
should hold an external reference to the event hub.
Flushing pending notifications can run arbitrary script, that can call Destroy()
on the pres shell (and thus tear down the accessible caret event hub, and the
manager with him).
I don't know why before my change this wasn't crashing badly, but the code as it
was just doesn't look sound to me at all either (maybe I'm misunderstanding
something and I should just revert that patch and give up on having nice
invariants during our flushes..., but I don't think it's the case).
This also adds some sanity-checking that we don't die under our flush.
MozReview-Commit-ID: 4s0UT0fD3TI
We're crashing because we're trying to flush layout of a shell whose document is
in the bfcache, so when we try to get the shell from the document in the style
set we get null back.
This shouldn't generally happen, and all flushes should go through the document
instead to avoid bugs like this and also other bugs related to subdocuments, and
all the stuff that nsIDocument::FlushPendingNotifications handles but the
presshell version doesn't. That's bug 1431852.
MozReview-Commit-ID: 1IJ4DBa3DCO
Fix warning: the parameter 'aHints' is copied for each invocation but only
used as a const reference; consider making it a const reference
MozReview-Commit-ID: 6CyT6gxVgES
Fix warning: use '= default' to define a trivial destructor
I added #include "nsDocShell.h" because I got an error "incomplete type
'nsDocShell' used in nested name specifier." After that, some files like
nsCanvasFrame.cpp fails to compile because they include
AccessibleCaretEventHub.h. Hence the moz.build changes.
MozReview-Commit-ID: BYZx7txvkSn
The issue is: when we hit some failure conditions in DragCaretInternal()
such as the frame is not selectable, or no frame is under the point, etc.,
we'll early return so that the auto scroll code is not being executed.
The logic in StartSelectionAutoScrollTimer() is similar to how
nsFrame::HandleDrag() handles the auto scrolling.
MozReview-Commit-ID: FtXZ8BWp3BX
The feature is controlled by pref layout.accessiblecaret.timeout_ms, and has been disabled in bug 1268410.
It's time to remove relevant code from the tree.
MozReview-Commit-ID: LLu8RiQcTpm
Override OnScrollPositionChanged() in ScrollState because we want to update
carets during scrolling in subframes without APZ.
Due to the observation in bug 1273045 comment 8, we do not distinguish
PositionChangedResult::NotChanged and PositionChangedResult::Changed.
Instead, we always update caret even if its position is not changed.
To avoid excessive CaretStateChangedEvents are dispatched in
OnScrollPositionChanged(), we add IsScrollStarted to distinguish whether
OnScrollStart() is called or not.
MozReview-Commit-ID: KNi9Mct4dSk
Part 2 is going to add a new hint which will use with RespectOldAppearance.
Hence this patch.
Remove #include "mozilla/WeakPtr.h" and "nsWeakReference.h" because they're
not used in the header.
MozReview-Commit-ID: KiNv0M0v8iO
In phoneRegex, replace '\\s ' (matching a whitespace character) with ' '
since phone number won't contain something like new line or tab.
Also, consider it done if selected text is not changed after calling
Modify().
MozReview-Commit-ID: 2lB9w2gYCOD
sHideCaretsForMouseInput is default to true on all platforms. When clicking
"SELECT ALL" on Fennec toolbar in <input>, AccessibleCaretManager will
receive SELECTALL_REASON with MOZ_SOURCE_TOUCH.
We should hide the carets only if the select-all reason is fired by
keyboard, not by touch.
MozReview-Commit-ID: D7FYIH5ZGS0
When the user switches back to a tab which has the selection highlight on
the document, call UpdateCarets() to bring the carets back.
MozReview-Commit-ID: LxNoNRl4FHZ
The mochitest for clipboard still needs the mozbrowsercaretstatechanged events,
so we disable "layout.accessiblecaret.hide_carets_for_mouse_input".
MozReview-Commit-ID: CD03lmjwUa9
This patch defines mozilla::SelectionType as an enum class. This is safer than nsISelectionController::SELECTION_* since setting illegal value to its variable is checked at build time. So, as far as possible, this should be used everywhere (but of course, this isn't available in scriptable interfaces).
And also this implements some useful methods for managing SelectionType and RawSelectionType which are implemented in layout/nsSelection.cpp because nsISelectionController is implemented by both PresShell and nsTextEditorState. Therefore, implementing one of them may make hard to find them. On the other hand, nsSelection.cpp is a better file name to look for them.
Note that this patch creates mozilla::Selection::RawType() for binding. Native code should keep using Selection::Type() but the binding code needs to use RawType() due to impossible to convert from SelectionType to RawSelectionType without explicit cast.
MozReview-Commit-ID: 81vX7A0hHQN
