Not really happy about the change but it addresses the problem.
* Loading the list from disk is much faster than parsing the string (2500ms vs. 30-50ms on a Nexus 6P)
* Because of (potential) disk I/O we are required to extract the label asynchronously on a background thread
* We load the list only once and then keep it in memory - like we did before the change.
MozReview-Commit-ID: 9MPGbmIGRnS
--HG--
extra : rebase_source : 8b31f852c6bb90fd57baeb07ba0066421d5a6e46
When a Gecko text change spans larger than our original request, we have
to do the replacement in parts in order to preserve any spans from the
original request. There was a bug where the selection is moved to the
wrong offset after the three replacements. This patch switches the order
of the two replacements and manually fixes up the selection.
For any text changes that originated on the Gecko side, this patch also
splits the replacement into two parts (delete + insert), so that old
composing spans are properly cleared first. This new behavior changes
the expected result for the test added by bug 1051556, so the test is
changed as well.. I think this new behavior is more correct, but if it
results in regressions, we can reevaluate.
Right now we send a "process-gecko-events" message to
GeckoInputConnection in order to wait on Gecko during testing. However,
now we have GeckoThread.waitOnGecko() to do that, so we can just use
that directly.
Under asynchronous IME, when we check text/selection for correctness, we
have to wait for the IC thread to sync the shadow text first. In order
to do that inside the assert methods, we have to move them to inside
InputConnectionTest, so that they can call processGeckoEvents() and
processInputConnectionEvents().
Also, ignore a single newline character, if present, at the end of the
actual text, because it's a side effect of some editing operations in
Gecko (e.g. clearing all text in an HTML editor).
We want to access ActivityStreamContextMenu directly in our new test. That menu
extends BottomSheetDialog which is contained in the design support library, hence
we need to make that library available in robocop builds in order to access
the AS context menu.
MozReview-Commit-ID: EPyv7wvkJAE
--HG--
extra : rebase_source : cafe1e7dd6fbdcfb471cf5f66a1c0522e796c3aa
Okay, this patch is much bigger than I anticipated. However at this point it is impossible
to separate all the interweaved changes.
* Update sizes and colors
* Use a dynamic number of top sites tiles to adjust to the screen size
* Do not keep references (to possible outdated) Cursors from every top sites page
* Remove unused bottom panel
* Add menu button to highlights items
MozReview-Commit-ID: 2CeEGCOXBKl
--HG--
extra : rebase_source : a780ec20fa6f87520c3418403ae4fe259ff39d69
Also make a temporary testing adjustment to handle the fact that TabsListLayout
and TabsGridLayout don't currently share a common usable base.
MozReview-Commit-ID: HPExAnehKRq
--HG--
extra : rebase_source : 4d9d0ce9fff216090ae676163e22645cd45c92a5
extra : source : 411a8ac7368e3b69e8191812fb53f0d1b8aa3247
It looks like the update (23.4) version of palette returns the actual dominant colour
for a red mock (FF0000). Previously it returned a more muted red (F80000).
MozReview-Commit-ID: 5VpRavyooHY
--HG--
extra : rebase_source : ab55e787a305481f890d00739de2843b87063bed
I am open to replacing "constellations" with a different name, but it was the best
I could come up with so far.
MozReview-Commit-ID: 9pljV8nC6O8
--HG--
extra : rebase_source : 781b5e1776adf457c92cacfecc09dc7afe03ad41
We need to ensure we can explicitely close any database connections we open in the TestLegacyLoader,
so we initialize a BrowserProvider instance so that we can then call .shutdown() on it once we're done,
which will close any open connections.
Additionally, we need to make sure to use correct authority when registering (using AUTHORITY_URI produces dual mapping
of providers, from content://org.fennec... and from org.fennec...), and wrap our providers in a DelegatingTestContentProvider,
which will append PARAM_IS_TEST param to wrapped URIs - when present this ensure we won't try to load a per-profile database.
MozReview-Commit-ID: LnzPhGOe6OY
--HG--
extra : rebase_source : 76eaadaa78bed7918b90012f8bdcc996dc3b2024
Compile geckoview sources into a separate gecko-view.jar file, apart
from gecko-browser.jar. This lets us separate JNI binding generation,
among other things.
Let's try to load from the legacy loader only if there's one icon left and
the other loads have failed. We will ignore the icon URL anyways and try to
receive the legacy icon URL from the database.
MozReview-Commit-ID: Kr7gHXBuAs7
--HG--
extra : rebase_source : 7fbdd507fa2c0a9aa4223db1da6aa5fbc1aa4907
If we are allowed to load the icon from the network then skip loading from the legacy storage and just
load a fresh icon. This will avoid touching the legacy storage (disk) every time before downloading an
icon.
MozReview-Commit-ID: C9hYqISno6U
--HG--
extra : rebase_source : 6f19839c38d37916deb351b3e080e023e532a83f
The FaviconGenerator is the last class in the old favicons package. As this class is only used by the IconGenerator
let's move the code inside the new class and remove FaviconGenerator.
MozReview-Commit-ID: 7NsJRGdoUWv
--HG--
extra : rebase_source : 703b80560f47220d094de0049cd9b92f8cd524eb
Most of the code in the favicons package is now obsolete and can be
removed. Most of this code still exists in the icons package (in smaller
chunks).
MozReview-Commit-ID: 6J4pgd5doNt
--HG--
extra : rebase_source : fa04223a758483d8440167fb63fe852ea3bde3f8
This patch does multiple things:
1) It restructures the icon code to follow a preparer, loader, processor pattern.
Instead of very long procedures we now have a lot of small components. This patch
includes 90+ tests for those components.
2) It replaces the database storage with the disk lru cache. We still keep the
tables around because we will still load from it as fallback to avoid needing
to migrate all data.
This patch is pretty big but a lot of it is moving code around and breaking it
into smaller chunks. A later commit will remove now obsolete components.
By creating a consistent mapping page URL -> icon URL -> icon data this change
fixes the linked bugs (bug 1269821 and bug 1271634).
MozReview-Commit-ID: 1nkrZn286Gv
--HG--
extra : rebase_source : e8d2902dc8553e43297e9e3f0fd96bf04f28ba3a
- Introduce a new BatchingUploader class to handle storing records and keep track of batches/payloads
- Refactor upload runnable and upload delegate into their own classes
- Introduce Last-Modified and X-I-U-S handling into non-batching mode
MozReview-Commit-ID: 3JLExwQvYzM
--HG--
extra : rebase_source : 4a6a2db595fe72473b033acd4f3180fa799319d4
Sebastian Kaspari
Andrzej Hunt
Phil Ringnalda
Jim Chen
Tom Klein
Sebastian Hengst
Ryan VanderMeulen
dlim@mozilla.com
Carsten "Tomcat" Book
Geoff Brown
Dylan Roeh
Grigory Kruglov
Jan Henning
Benjamin Dahse
Edouard Oger