After enabling diff view and selecting two snapshots, wait for TAKE_CENSUS_DIFF_END
action in the test before deleting all snapshots. Otherwise, TAKE_CENSUS_DIFF_END
might be dispatched after clearSnapshots toggled diff mode off.
MozReview-Commit-ID: HNmIVn1SC8c
--HG--
extra : rebase_source : 5f4e5541d07b7ab911773c197430857c29be7c6d
This patch forces the increment step to 0 when the type of the inplace-editor is
PLAIN_TEXT.
This type is used by:
- markup view for editing text and comment nodes
- rule-view to edit selectors
- projecteditor resource tree to create or rename a file
All other inplace editors will still benefit from the increment/decrement feature.
MozReview-Commit-ID: 8tZiFMwmWM3
--HG--
extra : rebase_source : 5e379848e9ba09073dd9775a080fd66654fe9c9e
Inplace editors for the layout view should use the CSS_VALUE content-type since they
allow to edit a CSS value, and not just plain text.
MozReview-Commit-ID: 9mQA1fSmXvD
--HG--
extra : rebase_source : 952e83cc28a4a218f7b88f7bd7bb3b132aa1bc37
As commented in the test, using createNewRuleViewProperty with "width:"
will actually trigger two ruleview-changed events :
- one when focusing the property-name input
- one when focusing the property value input (which is automatically done when hitting
return in this case, since we only entered a property name)
MozReview-Commit-ID: GjJnwDr3ON6
--HG--
extra : rebase_source : d4f22915dfdd65d55b23a09f797402944870f017
There can be multiple compartments within the same zone, only one of which is a
debuggee. In this scenario, CCWs from other compartments into the debuggee
compartment should be traced and treated as roots. Therefore, dealing with CCWs
at the JS::Zone level is incorrect, and this patch changes the granularity level
to JSCompartments. If you look at the callers and uses of the function, it makes
much more sense now.
Additionally, it renames `JS_TraceIncomingCCWs` to `JS::TraceIncomingCCWs`.
--HG--
rename : devtools/shared/heapsnapshot/tests/gtest/DoesCrossZoneBoundaries.cpp => devtools/shared/heapsnapshot/tests/gtest/DoesCrossCompartmentBoundaries.cpp
rename : devtools/shared/heapsnapshot/tests/gtest/DoesntCrossZoneBoundaries.cpp => devtools/shared/heapsnapshot/tests/gtest/DoesntCrossCompartmentBoundaries.cpp
When editing the style attribute, do not trigger autocomplete suggestions if the style attribute
was already closed by a set of matching double or single quotes.
Added a new test and moved common test functions to a dedicated helper.
MozReview-Commit-ID: 5AfkNXWvqTA
--HG--
rename : devtools/client/inspector/markup/test/browser_markup_css_completion_style_attribute.js => devtools/client/inspector/markup/test/browser_markup_css_completion_style_attribute_01.js
extra : rebase_source : 98101b99cde59dba808bc91465ddcebfd1518dda
When clearing all snapshots, if currently displaying a diff, dispatch the
diffing.toggleDiffing action.
MozReview-Commit-ID: 1DyPZ7tHC1X
--HG--
extra : rebase_source : 595d182023567c01fc14b7fb25fc65406296466d
There are few easy wins that greatly improve our perf with React:
* Add the PureRenderMixin to pure components that always render the same thing
given the same props and state. This implements React's shouldComponentUpdate
as a shallow equality check of the props and state, so render() will be
skipped if they match the previous values.
* Change action dispatch functions so they are only created once instead of on
each render(). If an outer component creates new functions for each render()
call, then the PureRenderMixin equality check fails for only the new function
instances, even though no data has changed.
MozReview-Commit-ID: D96X048nEF4
browser_scratchpad_contexts.js and browser_scratchpad_execute_print.js used to
access javascript properties on the content window object by using 'content.'
directly.
Now these tests use 'ContentTask.spawn' instead, which required the
runAsyncCallbackTests helper to be changed so it would use a Task and yield
on the promises returned by prepare and then.
MozReview-Commit-ID: Cgw2mRxPrP4
--HG--
extra : rebase_source : de57e5c7ab155ee48f537941e0db7cf7f82ed266
Bail out from the markup view onKeyDown handler if any modifier is currently
true. All shortcuts specified in this handler are intended to be used without
modifier, so for now this approach is fine.
Added a test checking the use case mentioned in Bug 1230325, with the S
shortcut. In order to write the test, had to create an additional method on
the test actor to be able to wait for events in the window of the content process.
MozReview-Commit-ID: 67icou0HkfA
--HG--
extra : rebase_source : e7235c487134505563cae3c5ac8bdabd42dc4363