The intermittent failure is caused by the innate drawback of the present test,
it doesn't be consistent with the behavior how we actually block the media.
Since we always block media implicit, it would be set in nsGlobalWindow's ctor.
We would never call blockMedia directly, so we can remove the function.
MozReview-Commit-ID: IjYJi5OHQ3X
If the tab was resumed before, it could start playing any autoplay media without user's
permission after session restore.
MozReview-Commit-ID: C3DHIIsLtJA
DevTools are moving out of mozilla-central and m-c code should only rely on the
DevToolsShim that will remain available in m-c after the migration.
This changeset updates SessionStore.jsm to call the DevToolsShim instead of the
scratchpadmanager.
The actual underlying implementation is the same when devtools are installed and
no-ops when they are not.
MozReview-Commit-ID: Fv22TtjJiDw
This change means that any related http pages driven through content (window.open, links, etc.) will continue to be loaded in the file content process.
Same-origin loads via the UI will also remain in the file content process.
Cross-origin loads via the UI will cause a process switch.
History navigation will stay in the process, if it was originally loaded in that process.
If the tab was resumed before, it could start playing any autoplay media without user's
permission after session restore.
MozReview-Commit-ID: C3DHIIsLtJA
While setTab does end up setting the label and icon, it also does completely different and more crucial things. This call really doesn't belong in updateTabLabelAndIcon.
MozReview-Commit-ID: 1ZSY17ThNx
While setTab does end up setting the label and icon, it also does completely different and more crucial things. This call really doesn't belong in updateTabLabelAndIcon.
MozReview-Commit-ID: 9HXYS0fXgRN
restoreWindow only opens new tabs rather than re-use pre-existing tabs, so
there's no point in calling showTab anymore. hideTab also will not hide the
selected tab, so the "all tabs to be restored are hidden" case can't happen.
MozReview-Commit-ID: FzmRcjprkuN
This patch replaces FX_SESSION_RESTORE_CONTENT_COLLECT_DATA_LONGEST_OP_MS with
a more fine-grained data collection via a keyed histogram so that we can
determine which type of data collection we spend a lot of time on.
Additionally, it abandons sending telemetry data from the content to the parent
via custom messages and instead uses the Telemetry service directly. Thus it
also gets rid of a bug that currently overrides measurements for the same
histogram done quickly in succession
This patch replaces FX_SESSION_RESTORE_CONTENT_COLLECT_DATA_LONGEST_OP_MS with
a more fine-grained data collection via a keyed histogram so that we can
determine which type of data collection we spend a lot of time on.
Additionally, it abandons sending telemetry data from the content to the parent
via custom messages and instead uses the Telemetry service directly. Thus it
also gets rid of a bug that currently overrides measurements for the same
histogram done quickly in succession.
Bug 588482 made it possible to restore pinned tabs in windows before restoring the rest of the
tabs. Part of this involves extracting the pinned tab data from the last session, and make it
the default state to restore. The rest can be restored later.
If one of the pinned tabs in a window were selected, the code that was preparing the default
state for that window set the selection value by calculating the length of the tabs array
(before adding the selected one), and adding 2. Presumably, 1 count is because of the tab
we're about to add, and another count is due to how the "selected" value in SessionStore
is 1-indexed (where "0" means, "don't change the selection").
This is an off-by-one error though, since the tabs array length already cancels out the
need for the extra "1-index" difference. It is sufficient to just add 1 to the length of
the tabs array to calculate the selected value.
This appears to have been a bug for a while, but was covered up by the fact that the old
tab selection code exited early if the 0-indexed value of the selected tab was not within
the length of the tabs array. Bug 1351677 uncovered this bug by removing that second check.
This patch fixes the off-by-one error, and also puts a check back in to ensure that the tab
we're going to select is at a valid index.
MozReview-Commit-ID: 9WbbU0vUJHG
FX_SESSION_RESTORE_COLLECT_DATA_LONGEST_OP_MS can go because that's exactly the same as FX_SESSION_RESTORE_COLLECT_DATA_MS now.
We can remove FX_SESSION_RESTORE_COLLECT_COOKIES_MS because that's just a flat line since bug 912717 landed.
