`Selection::Extend()` is too slow but editor and ContentEventHandler use it in
some places. We should make them use `Selection::SetStartAndEndInLimiter()` or
`Selection::SetBaseAndExtentInLimiter()`. The former is usable only when caller
guarantees the start point is prior to the end point in the DOM tree.
Otherwise, we need to use the latter even though it's slower than the former.
Differential Revision: https://phabricator.services.mozilla.com/D23462
--HG--
extra : moz-landing-system : lando
`Selection::Extend()` is too slow but editor and ContentEventHandler use it in
some places. We should make them use `Selection::SetStartAndEndInLimiter()` or
`Selection::SetBaseAndExtentInLimiter()`. The former is usable only when caller
guarantees the start point is prior to the end point in the DOM tree.
Otherwise, we need to use the latter even though it's slower than the former.
Differential Revision: https://phabricator.services.mozilla.com/D23462
--HG--
extra : moz-landing-system : lando
Currently, IMEStateManager::OnChangeFocusInternal() tries to sync the state
whether menu keyboard listener is installed between itself and active remote
process -- When menu keyboard listener is installed, it posts a message to
_only_ active remote process. When menu keyboard listener is uninstalled,
it posts a message to _only_ active remote process. So, it's not guaranteed
that active remote process at installing and uninstalling may be different.
If it's different, IMEStateManager in the old remote process believes that
menu keyboard listener is still installed. This is what the cause of IME
unavailable in a remote process.
Current approach must be wrong. IMEStateManager should manage menu keyboard
listener state only in the process which the listener is installed in. Then,
when menu keyboard listener is uninstalled, IMEStateManager needs to restore
the latest input context in the remote process without asking the remote
process.
Therefore, this patch does:
* stops IMEStateManager::OnChangeFocusInternal() posting message when menu
keyboard listener is installed and uninstalled.
* removes the message sender and receiver from PBrowser.
* cache the latest input context of active remote process in
IMEStateManager::SetInputContextForChildProcess().
* make IMEStateManager::SetInputContextForChildProcess() not set input context
when menu keyboard listener is installed in the process.
* tries to restore latest input context in the remote process in
IMEStateManager::OnChangeFocusInternal(). If there is no cached input
context, it does nothing and waits next SetInputContextForChildProcess() call.
* clears the cache when IMEStateManager::OnChangeFocusInternal() changes
active remote process to different one or nullptr.
So, this must improve performance at activating and inactivating memubar and
opening and closing popup menu in the main process.
MozReview-Commit-ID: EelKSPlaXdw
--HG--
extra : rebase_source : db7334b3c0d3ce87868450ee3179692027975bd6
Android now supports telling an IME that it shouldn't store user-entered content into it's dictionary/language model/etc. and we want to automatically enable this in private browsing.
As the code that handles input on Android doesn't have any notion of tabs (and therefore of the difference between normal and private tabs), the best way to get that info across is to retrieve it directly within the IMEStateManager from the corresponding document and store it in the inputContext, which is then passed to Java for Fennec to handle.
Implementing this within Gecko also has the benefit that this part of the code can be used by other platforms as well should they want to support similar features in the future.
MozReview-Commit-ID: DsxjC4Ma7DR
--HG--
extra : rebase_source : f0940cb40170ab32cf5a172d07a61d083427be8a
Currently, IMEStateManager::OnChangeFocusInternal() sends blur notification to IME when a remote process has IME focus and focus is moving from the process. However, if IME wants to keep composition even during deactive and nobody will gets focus (i.e., all windows becomes deactive), IMEStateManager shouldn't send the blur notification because it causes committing composition.
Therefore, it should send blur notification only when focus is moving to a PresContext (that means that not all windows becomes deactive) or IME doesn't want to keep composition during deactive.
Then, even if another window becomes active next time, IMEStateManager can send "stop IME state management" message to the composing remote process and the remote process can commit composition normally.
Additionally, this patch ensures to send blur notification when IME focused TabParent or widget is being destroyed. This fixes new memory leak bug of this patch (sFocusedIMETabParent keeps grabbing the instance until shutting down in some mochitests).
MozReview-Commit-ID: KYiFGo970a8
--HG--
extra : rebase_source : c3de0df442420979414b47d8d20c7988c49b205b
Requests to commit/cancel composition came from remote process with sync message. So, it may be too late. E.g.,
* If the process already sent new composition start but is not handled by the remote process yet.
* If the process already send commit message but it's not handled by the remote process yet.
* If focus was already moved to different process.
In the former 2 cases, the remote process should wait eCompositionCommit(AsIs) events for clearing TextComposition. Therefore, the requested should be treated as it's handled asynchronously.
In the last case, the remote process should commit composition with latest composition string in the main process because if the remote process commits composition with "current" composition string in it, user may lost some inputted text.
MozReview-Commit-ID: 18BUoZZq7HS
--HG--
extra : source : fd1585ad670a87d8b1ef8908931f3d4037751475
IME should receive notifications and requests only from proper process. E.g., IME shouldn't commit composition by a request which came from previous focused process.
This patch makes that IMEStateManager::NotifyIME() takes pointer to TabParent optionally. If the request or notification came from remote process, it should be non-nullptr. Then, this makes it ignore notifications and requests from unexpected process.
Note that this patch also touches some gfx headers because they use |ipc::| but compiler is confused at the ambiguousness between |mozilla::ipc::| and |mozilla::dom::ipc::|.
Finally, this patch changes the NS_ASSERTION in IMEHandler::OnDestroyWindow() to MOZ_ASSERT because the orange caused by the NS_ASSERTION was not realized since there was already an intermittent orange bug caused by different NS_ASSERTION.
MozReview-Commit-ID: 9CgKXQRJWmN
--HG--
extra : source : f3b5711908870c5e0e852a399a07e0ae721a12f1
Currently, IMEStateManager always sets input context as set by current process even when it needs to adjust IME state when a tab parent for current focused IME process is removed. Then, input context for the widget is marked as for main process but the widget still have IME focus of a remote process.
For fixing this mismatch, IMEStateManager should set ORIGIN_CONTENT even when the tab parent is being destroyed.
MozReview-Commit-ID: C10YOAtkET4
--HG--
extra : source : 9430d123b19e0ac551c6048bb044fcfa22d13e45
IMEStateManager should cache nsIWidget for sPresContext at caching sPresContext. Then, even if sPresContext has gone, IMEStateManager can clean up with the nsIWidget cache.
Unfortunately, editor has some bugs about calling IMEStateManager::UpdateIMEState(). That is, calling it *before* IMEStateManager::OnFocusChange(). In such case, this patch makes UpdateIMEState() ignore the call.
MozReview-Commit-ID: 1cydI03WyB8
For sending NOTIFY_IME_OF_COMPOSITION_EVENT_HANDLED after the other change notifications which was caused by the user input, we need to use IMEContentObserver::IMENotificationSender because it sends the notifications when it's safe to do it.
This patch makes TextComposition use IMEContentObserver to send the notification. However, if there is no active IMEContentObserver, e.g., composition events are fired on unfocused window, TextComposition sends it by itself (same as current implementation).
If IMEContentObserver stops observing when it has pending NOTIFY_IME_OF_COMPOSITION_EVENT_HANDLED, it cannot send the notification (i.e., it is discarded completely in such case). However, in such case, IMEContentObserver sends NOTIFY_IME_OF_BLUR. So, anyway, native IME handler should treat the blur notification as it including NOTIFY_IME_OF_COMPOSITION_EVENT_HANDLED.
On the other hand, we're buggy if composition events are fired in non-active window. Even in such case, IMEContentObserver should be created for active editor in each document and it notifies IME of the changes. But this is out of the scope of this bug.
MozReview-Commit-ID: 7Q0ZsJTh4hX
--HG--
extra : rebase_source : 6417f991fa8c0fbe3f25b27bacf4257e5485aecc
PuppetWidget::GetInputContext() needs to communicate with its parent process with synchronous IPC. This is very expensive for focus move.
Currently, IMEStateManager uses nsIWidget::GetInputContext() only for checking the IME enabled state. Therefore, it's enough to cache input context when nsIWidget::SetInputContext() is called. Then, we can avoid to communicate with synchronous IPC with PuppetWidget::GetInputContext() in most cases.
This patch makes IMEStateManager stores the last widget which sets input context. When PuppetWidget uses its input context cache, it should check if it is the last widget to set input context with IMEStateManager since an input context may be shared with other widgets and another one may have update the input context. I.e., PuppetWidget's input context cache may be already outdated after IMEStateManager sets input context with another widget.
This patch gives up to support retrieving IME open state from child process. However, perhaps, this is not necessary for everybody including add-on developers because the only user of IME open state in child process is nsIDOMWindowUtils. So, add-ons can send IME open state from chrome process instead. If this decision is wrong, unfortunately, we should support it again in another bug. It's easy to support with creating another nsIWidget::GetInputContext() or adding additional argument to it.
MozReview-Commit-ID: B2d2CCTsPKj
--HG--
extra : rebase_source : 4117330ba7871753176da960063b612e96f11752
Henri Sivonen
Peter Van der Beken
Ryan Hunt
Masayuki Nakano
Noemi Erli
Sylvestre Ledru
Boris Zbarsky
Jan Henning
Jim Mathies
Wes Kocher
btian
Randell Jesup
Robert O'Callahan
Jim Chen
Phil Ringnalda
Andrew McCreight