Convert events used in PageActions to bundle events. UI thread events
are used because the listeners require the UI thread. The observer
notifications from Java to Gecko are converted to events as well.
Convert events used in AccountsHelper to bundle events. Most events are
kept as Gecko thread events, but a couple events that start activities
are converted to UI thread events.
Convert events used in IntentHelper to bundle events. UI thread events
are used for most events because the listeners perform operations on
Intent objects, which is best done on the UI thread. For
"Intent:GetHandlers", use a Gecko thread event because it's possible we
want a synchronous callback response.
Convert events used in HomePanelsManager to bundle events. Background
thread events are used because HomePanelsManager processes panel changes
in the background thread. Changing to background thread events also lets
us make the change queue a simple ArrayList instead of a
ConcurrentLinkedQueue, because there is no longer multiple threads
involved.
Convert the "HomePanels:Data" event to a GeckoBundle/BundleEventListener
event. UI thread event is used because the listener invokes the callback
on the UI thread.
Convert events used in FindInPageBar to GeckoBundle/BundleEventListener
events. UI thread events are used because the listener performs UI
operations. FindInPageBar also sends some events like "FindInPage:Find"
from Java to Gecko; those events will be converted in another bug.
Convert observers in ActionBarHandler.js to events that go through
WindowEventDispatcher. "TextSelection:Get" now replies through the
callback interface rather than a "TextSelection:Data" event. The patch
also adds new lazy-loading events in browser.js to replace the
lazy-loading observers that ActionBarHandler relied on.
Convert the events used in MediaCastingBar and ChromeCastPlayer to
GeckoBundle/BundleEventListener events. UI thread events are used
because the listener performs operations on the UI thread.
When going down one folder level, we store the current scroll position of the list view and then scroll to the top of the list, so that we always show a child folder starting from the beginning. When navigating back up to its parent, we then restore the previously stored scroll position.
MozReview-Commit-ID: 9C2RMXrlUm1
--HG--
extra : rebase_source : 7ec5a3f1d602db9454efa52603e0f50704c1418d
Fix up javadoc so it matches the actual function parameter and remove unneeded imports.
MozReview-Commit-ID: I91NuqLlpL1
--HG--
extra : rebase_source : 13366049ee9fe1487b3cf24f4e6f610f66bee5ee
Showing the tab history panel involves a gecko call to retrieve tab history. This
can be slow, meaning we have no idea what state the app will be in when the tab
history data is returned. Thus we need to protect the code that shows the tab
history fragment against a number of scenarios to avoid crashes in those cases
where the app might be shutting down:
1. If onSaveInstanceState() has been called (which might happen before or after onPause(),
and might be linked to app shutdown - but the docs don't appear to give any guarantees),
fragment transactions cannot be performed. We protect against this by accepting loss
of state in fragment transactions.
2. If the Activity has been completely destroyed, trying to perform a fragment transaction
will likewise fail. We protect against this by not even trying to perform the transaction
if we definitively know that the Activity is being shut down (ie isFinishing()).
In both of these cases, we simply must accept that we're potentially losing UI state: i.e.
a user could request the tab history panel via long-back-press, followed by exiting the app;
we now end up never ever showing the panel. This scenario doesn't seem like a major loss - and
fixing this issue properly would require significant investment (i.e. we would need to either
cache tab history on frontend side, or cache the tab-history panel request - and it's not clear
users will still care about seeing the panel the next time they open firefox).
MozReview-Commit-ID: JsAK1By8yqn
--HG--
extra : rebase_source : 9a3b64d2829a969e4de2775c2714874aa8708031
Since bug 1228593, the mobile session store
- once again flushes its data when we are quitting, to make sure the latest state (including any potential cleaning of history/tabs) is flushed to disk
- ignores windows/tabs closing as a byproduct of shutdown
The latter point is dependent on a new shutdown notification introduced in that bug. Because we forgot to add that notification to the restart code used for add-on updates, in that case the session store currently doesn't enter shutdown mode and therefore records the window being closed during shutdown before flushing its data to disk, which means that all open tabs are lost.
MozReview-Commit-ID: LgtdQoYwacM
--HG--
extra : rebase_source : 1bf6c544d9e25961a0e64236678ca5938e8a69fe
A tab object should know how to zombify itself instead of having to rely on the MemoryObserver. This also simplifies the situation for anybody else who wants to call this function since it is no longer necessary to figure out how to load the MemoryObserver for this.
MozReview-Commit-ID: 5IX114QUjBT
--HG--
extra : rebase_source : 964186678848b804e5f0550e1b6a328d7fcfbc50
Actors outside of the session store shouldn't have to poke around within the session store's data structure and end up reimplementing parts of the session store code (and cause data loss if the implementation is incomplete) in order to restore delay loaded zombie tabs. Therefore, we simply expose a method for this via BrowserApp's tab object.
To simplify handling and make the method a little more fool-proof for external callers, the check whether the tab is actually zombified is moved into the restoreZombieTab() function.
Later on, we can also hook up this method to the appropriate web extension API (see bug 1322485).
MozReview-Commit-ID: 85lnbCpMcP3
--HG--
extra : rebase_source : a6f1cfa11debcb18471b49804776521c60655fce
With --with-intl-api, xpcshell test for MozIntl is always failure on Android. So we should add mozIntl.xpt to package.
MozReview-Commit-ID: BWBKSG3gSqr
--HG--
extra : histedit_source : d4091b91221bc2ef282d0855dbf4856adb403c9b
Bug 1319660 - 1. Don't take shortcut if old replacement ranges don't match; r=esawin
The block at [1] is a shortcut we take when we reconcile Java text
changes with Gecko text changes. However, we only checked that the new
ranges are the same, i.e. that the new Gecko text is the same as the new
Java text. We should also be checking that the old ranges are the same,
i.e. that the replaced Gecko text is the same as the replaced Java text.
[1] https://dxr.mozilla.org/mozilla-central/rev/bbbd2f7539f224a482cc6d2dd10e6a5f31c8baf3/mobile/android/geckoview/src/main/java/org/mozilla/gecko/GeckoEditable.java#1233
Bug 1319660 - 2. Use previous node instead of sibling when adjusting last node; r=masayuki r=smaug
nsContentIterator in pre mode adjusts its last node if the node is a
childless node like <br>. However, right now it's using GetPrevSibling,
which can lead to error in some edge cases such as:
<p></p><div><br></div>
In this case, if the last node is <br> with offset 0, GetPrevSibling
will return <p> because <p> is <br>'s parent's previous sibling, and the
last node will be set to <p>. However, the correct last node in this
case is <div>, because <br> with offset 0 refers to the position to the
left of <br>, which is <div> with offset 0. In this case, PrevNode
returns the correct <div> value, so we should set the last node to the
result of PrevNode.
For the first node, for a childless node in pre mode, GetNextSibling and
NextNode are the same, so there is no bug in this case. Nevertheless,
this patch changes the call to NextNode to be consistent with calling
PrevNode for the last node.
Bug 1319660 - 3. Add test for correctly adjusting last node in content iterator; r=masayuki
Add a test for the previous patch that makes sure querying selected text
in an edge case works correctly.
Bug 1319660 - 4. Add test for start node regression; r=me
Add a new test case for the NextNode() regression. r=me for trivial
test-only patch.
Bug 1319660 - 5. Restore GetNextSibling call for first node of pre-content-iterator; r=smaug
The last patch changed the `GetNextSibling()` call to `NextNode()`
because I assumed they're equivalent in this case. That turned out to
not be the case because we can reach this line even if the node has
children -- the index just has to be after the last child. So this patch
restores the `GetNextSibling` call to restore the correct behavior.
I also added some comment to clarify that we can reach this line due to
one of two conditions: 1) the node has no children; 2) the node has
children but the index is after the last child.
This patch also replaces the `HasChildren()` check when setting
`cChild`. If the index is after the last child (i.e. index ==
childCount), `GetChildAt()` fails and we erroneously log an assertion
warning, even though the input was valid. The new check handles all
cases whether start node has children or not.
Bump the Android builders to the latest beta release to reduce
the variance when we update Firefox 53 to 1.15.0 stable early
in the Aurora phase.
Android builds were moved to 1.15 early to address a code generation
issue with devices without neon.
Work around an issue with tarball naming in the cargo packages.
MozReview-Commit-ID: KQfkWmXV9hQ
--HG--
extra : rebase_source : 9448e0b948740fc3905ef70c8df316dc7342d52e
When turning off SPS profiler by configure option, or we build non-SPS arch build such as android/aarch64, Tab.prototype.onStateChange already throws the exception because Profiler isn't defined.
So we should check AppConstants.MOZ_ENABLE_PROFILER_SPS to use Profiler object.
MozReview-Commit-ID: A9ISurxiRmc
--HG--
extra : rebase_source : 25103e97cd4827edef33335e3aec9384a6695526
When a snackbar with a button callback is dismissed, this translates to a rejected promise from sendRequestForResult(). We need to catch this in order to avoid a spurious 'JavaScript Error: "uncaught exception: undefined"' message appearing in the console and possibly causing confusion.
MozReview-Commit-ID: 7hsAOAMTeDP
--HG--
extra : rebase_source : 6c5eb28d2e0dcf39a35b310d1e1c45cfc47f272b
We've been parsing the session store file and doing the initial tab creation on the Java side for a long time now, so the code to open completely new tabs on the session store side instead of using tab stubs is no longer necessary.
Besides, because of bug 1301160 the on-disk session file is now containing tab IDs as well, so this distinction between tabs read directly from the raw session file and those contained in a session string passed over from the Java UI is no longer available.
MozReview-Commit-ID: Hmq0TwClqMQ
--HG--
extra : rebase_source : 1f9ab6da18a7fd0f40a1cb0784082abb2384404f
This patch adds the ability to run SingletonEventManager handlers in
different modes: sync, async, raw (no exception handling, arg cloning,
or asynchrony), or asyncWithoutClone. When you call the handler,
you're required to specify which variant you want.
Existing uses of SingletonEventManager are all converted to async calls.
Note that some of them were previously synchronous, but it didn't appear
to be necessary.
Also added a callOnClose for SingletonEventManager when the last listener
is removed.
MozReview-Commit-ID: ATHO97dWf3X
--HG--
extra : rebase_source : 43988ee453e965e1c591b1c64603c22f5a103cff
The protocol handler is intentionally simple. It works for the entering into the
urlbar case but not a lot else. Included some basic tests.
MozReview-Commit-ID: 4FQ1irdt3Nj
--HG--
extra : rebase_source : 30d9fc7ba2c500f268d282e47d291724d9899ba4
Unfortunately tinting the bookmark star is highly complicated due to the tinting
that NavigationView performs (i.e. we'd likely have to disable NavigationView
tinting, and do manual tinting on every icon - alternatively we could hack
the tint-list to use blue for "checked" items, and set the bookmark item
as checked). Since it's unclear if we even want the star to be blue,
we'll leave it grey (but filled) for now.
MozReview-Commit-ID: DekRZJayKIz
--HG--
extra : rebase_source : 1e3443e5f09c60bd0d7295bc35ecc08ca17b3dab
See https://code.google.com/p/android/issues/detail?id=209558 . On Devices
running Android 4 and below, VectorDrawable's can be corrupted due
to overzealous proguarding. This doesn't appear to have been
fixed in the support library yet, and even if it were fixed we
still wouldn't be able to switch to the most modern support
library without significant work.
MozReview-Commit-ID: 3ByogGygCEd
--HG--
extra : rebase_source : 2adeed63f88ca39a71feec60627fb812b76d3bb4
Previously if we could not compare two icon descriptions we'd always return the "right" one. This does
not create a consistent order if the parameters are flipped. As a result some operations on a TreeSet
can fail (like remove()).
MozReview-Commit-ID: EYPlhzGUEnD
--HG--
extra : rebase_source : f8918c022188401e21a03ac666628cff3e87f317
Also refactor MediaControlService to be test-friendly.
MozReview-Commit-ID: 344cEEuVroG
--HG--
extra : rebase_source : 14042bc3225538f621605b7aadb9d729c01ac62b
When the playing-media changes (ie. in Youtube click another
video when it is alreay playing one), there will be one more
event MEDIA_PLAYING_CHANGE, and the tab is still playing media.
This fix is to handle this situation then update state.
MozReview-Commit-ID: Aje2BBqFZKE
--HG--
extra : rebase_source : f7c8f80c1d49e3a9aef868d11a4cd9d722a5c221
This patch involve State into MediaControlService.
ACTION is behaviour for an intention, and so far we use it for storing
media state. Now involve enum to distinguish State from Action.
So flow becomes: Action happens -> Change State -> UI reflect state
changing.
MozReview-Commit-ID: FBHn02tgVO6
--HG--
extra : rebase_source : e1c57502a13669fad0ee2848c6cff420ebbe064f
Convert the "Tab:*" observers in browser.js, which go through the observer
service, to events that go through GlobalEventDispatcher.
FindHelper.js uses "Tab:Selected" through a lazy loader, but I don't
think FindHelper should listen to "Tab:Selected" at all until it is
being used. THerefore, this patch adds "Tab:Selected" registration to
FindHelper.js itself.
Convert events used in Tabs to GeckoBundle/BundleEventListener events.
All of the events are converted to UI thread events, because a lot of
the listeners modify states in Tab (e.g. through
`Tab.handleLocationChange()`), and those states are later accessed on
the UI thread. We used to modify the states on the Gecko thread, and
then access them on the UI thread, which introduces potential race
conditions. Therefore, I think it's best to convert these events
wholesale to UI thread events to avoid races.
Tabs.notifyListeners now calls the listeners directly if it's already on
the UI thread, instead of posting to the UI thread again. This is meant
as an optimization because the events are now coming in on the UI thread
already.
The "Tab:SelectAndForeground" and "Tab:Select" events are merged into
one "Tab:Select" event, and a "foreground" option is added.
The "Tab:StreamStart" and "Tab:StreamStop" events are merged into one
"Tab:RecordingChange" event, and a "recording" option is added to
indicate starting or stopping.
Make SiteIdentity.update use GeckoBundle, and also simplify the enums
in SiteIdentity, by using Enum.valueOf to get enum values from names,
instead of custom code.
Convert events used in ZoomedView to GeckoBundle/BundleEventListener
events.
"Window:Resize", "Browser:ZoomToPageWidth", and
"Browser:ZoomToRect" listeners are removed because they are not sent
anywhere.
The "Content:LocationChange" listener is merged into the
"Gesture:CloseZoomedView" listener, so that we can leave
"Content:LocationChange" alone for now. Otherwise we would have to
convert the "Content:LocationChange" listener in Tabs as well.
The listeners are changed from global listeners to per-GeckoApp
listeners, because two of the events, "FormAssist:AutoCompleteResult" and
"FormAssist:Hide", are per-GeckoApp events also used by FormAssistPopup.
Because the events are now per-GeckoApp events, they are now registered
from onAttachedToWindow and onDetachedFromWindow.
Convert "FormAssist:*" events in FormAssistPopup to
GeckoBundle/BundleEventListener events. UI thread events are used
because the listener performs operations on the UI thread.
Also convert the "FormAssist:*" observers in browser.js to use
WindowEventDispatcher.
The "FormAssist:AutoComplete" event that goes from Gecko to Java is
renamed to "FormAssist:AutoCompleteResult" to prevent conflict with the
event with the same name that goes from Java to Gecko.
While here, fix abuses of GRE_MILESTONE in moz.build and use
RELEASE_OR_BETA instead.
--HG--
extra : rebase_source : 2a8c6501b0fe3bb603a65d52dda13c609a75b033
This is relevant when a new tab is opened (e.g. through an intent from an external app) just as the user has pressed the Quit button (see bug 1266594), in which case the TabOpen event might reach the session store after Gecko shutdown has already started. If we're still in the preliminary phase and the browser window hasn't yet been closed, we can persist this new tab by collecting the session store data from the browser window once more.
Additionally, we now reset the time of the last backup when clearing the session store data, so the backup file is recreated as soon as possible after a history/tabs sanitisation, especially for users who are cleaning private data on exiting Firefox.
MozReview-Commit-ID: 9BmqrSIUJ42
--HG--
extra : rebase_source : be3c413db16523aac69d8283f36057713c1e9367
