Currently, Gecko decision tasks spend ~8 minutes making HTTP requests to the
queue API. Let's throw a thread pool at it so tasks are submitted faster.
This could possibly be faster if there were a single batch submit API in
the queue. However, that isn't implemented for technical reasons described
in the bug.
MozReview-Commit-ID: MeHItEVBbk
--HG--
extra : rebase_source : ebe182c5906df70cd2e4cbd545f7b7b9189882da
Although I haven't been able to pinpoint why, it looks like
nsPerformanceStatsService::Dispose() may be called twice, which in
turn causes crashes. This patch makes sure that calling the method
twice is idempotent.
Also, just in case this was due to a typo in
AddObserver/RemoveObserver, this patch replaces the literal strings
used in both with constants.
MozReview-Commit-ID: 8fXO20r5xvO
--HG--
extra : rebase_source : 490f1a5186a426a41ab567e564cdbd46080262ec
For people with customised home panels, we need to explicitly remove the Recent Tabs panel. We also unhide the Combined History panel if it was previously hidden and additionally turn it into the default panel if the Recent Tabs panel was the previous default panel.
MozReview-Commit-ID: 5CSJUTRysQU
--HG--
extra : transplant_source : A.%23%F5%5D%CE%09%E4%F1%BE%24%7E%13Q%1B%0Bx%15e%91
By passing the panel types to be removed/set as new default panel as arguments instead of hard coding them, we can reuse that function for our own home panel config migration.
MozReview-Commit-ID: BsMxcbInRbX
--HG--
extra : transplant_source : %89zh6%85%F9r%D5%BDu%3E%80%86%AB%7C%1D%B6P%8E%0B
Hiding the header was causing some problems so instead restructure the
XUL to put the discovery pane in the top-level deck alongside a vbox
with the header and a deck with all the other panes that have the header.
MozReview-Commit-ID: 2PnW1F9aYgt
--HG--
extra : rebase_source : 758e8560c88801e65ae7e0d1152bd5b75e013b70
Once we have references to all the atoms for the filters list, we can
pass all those references in, rather than having to copy the array and
thus the references.
I didn't include nalexander's MOZ_INSTALL_TRACKING suggestion because my make
skills are too shaky to make this worthwhile. Specifying a keyfile when
MOZ_INSTALL_TRACKING is disabled isn't harmful afaik (though it's a little
spammy).
Also, added code comment duplicated for emphasis:
# (bug 1277553) In Aurora -> Beta simulation builds, no update channel is
# specified, causing an assertion to throw that MOZ_INSTALL_TRACKING is
# specified but the keyfile is not. In this case, we add a default keyfile.
# This has the disadvantage that if our beta/release checks above ever
# fail, we'll come to this default case and the compile-time check to
# specify a valid keyfile will be broken. I don't have any better
# alternatives.
MozReview-Commit-ID: 7tmemvpDaW8
--HG--
extra : rebase_source : 96930d595ebc22c06dadea1d28782ec6a2c023c2
Additionally, we add this file to the tree so it can be used by bug 1277553.
The following commit will add the docs explaining how to use the file I
added.
The added file contents are slurped directly into
AdjustConstants.MOZ_INSTALL_TRACKING_ADJUST_SDK_APP_TOKEN. We get the sandbox
token from the previously working value in [1].
An alternative solution would be to remove the assertion that
`--with-adjust-sdk-keyfile` is specified and provide a default value if the key
is not specified. However, after investigating bug 1277553, I dislike this
approach because we lose the compile-time checks that a key file is specified,
which is dangerous for release/beta builds where the key file may not be valid,
we push a release, and we don't get the data we expect.
Ideally, we specify `--with-adjust-sdk-keyfile` for non-MOZILLA_OFFICIAL
builds, but I don't know how to do that.
[1]: https://hg.mozilla.org/mozilla-central/rev/ac75f8f4c01d#l1.26
MozReview-Commit-ID: HPNgiwfzU5p
--HG--
extra : rebase_source : 6ee6cf869624b2274338d76b63072d13cc3c8432
Existing keyboard shortcuts have been reimplemented using the new shortcut-key
api.
A keypress handler was removed on the breadcrumbs buttons, it had no use
since a focused button is always selected.
MozReview-Commit-ID: JTSRxwQGSlh
--HG--
extra : rebase_source : 4466162a2cf2bd77cb50ee77ea5843c973ceb5b2
aOutputLeft is null checked and then dereferenced twice more in this function,
and all callers pass a non-null pointer. So just remove the null check.
--HG--
extra : rebase_source : 8190a51fde8434ac346a4e23db5ed4703762778c