This patch exposes the objectInspector reducer as a regular
file that can be required in the devtools codebase.
This should make things faster to start as we won't need
to load the entire bundle (especially for the inspector where
the reps is only used when navigating to specific subpanels).
This means we need to remove the flow typing for now since
we don't have a cross-toolbox build process like we have
in the debugger.
Differential Revision: https://phabricator.services.mozilla.com/D47728
--HG--
extra : moz-landing-system : lando
This patch exposes the objectInspector reducer as a regular
file that can be required in the devtools codebase.
This should make things faster to start as we won't need
to load the entire bundle (especially for the inspector where
the reps is only used when navigating to specific subpanels).
This means we need to remove the flow typing for now since
we don't have a cross-toolbox build process like we have
in the debugger.
Differential Revision: https://phabricator.services.mozilla.com/D47728
--HG--
extra : moz-landing-system : lando
- Use the shared error.svg icon for security errors.
- Make the Netmonitor's padlock icons accept colors from CSS, and use the same colors as Console for errors and warnings.
- Turn all security state icons white in selected rows.
- Tweak some icons, e.g. make the shield icon's strokes a full 1px instead a slightly thinner strokes which made
Differential Revision: https://phabricator.services.mozilla.com/D47560
--HG--
extra : moz-landing-system : lando
We weren't using the shared defer function because
at some point it was still using Promise.jsm,
which is slow.
But since Bug 1388054, it's not using Promise.jsm anymore,
so we can use the shared function.
We take this as an opportunity to clean up
the shared defer file comments.
Differential Revision: https://phabricator.services.mozilla.com/D47739
--HG--
extra : moz-landing-system : lando
When clicking the invoke button on a property that
needs to be quoted, e.g. hello-world, the result
was undefined.
That's because we were sending the property name to
the server with the surrounding quotes.
This patch fixes the issue and adds a couple of test
case to make sure we don't regress.
Differential Revision: https://phabricator.services.mozilla.com/D47399
--HG--
extra : moz-landing-system : lando
The disableTask options is renamed to enableTaskMiddleware,
which defaults to false.
This caused failure in the dom mutation breakpoint test, because
we weren't waiting until the call to the server to remove the
breakpoints were done.
It wasn't an issue before because this was handle by the task
middleware, which catch rejections.
We fix this by only dispatching the action when the breakpoints are
indeed removed. We also tweak to waitForAllElements helper to be
able to wait for a specific count of element (it used to resolve
when there was at least the specified count, which is not ideal
when trying to assert removal of an element).
Differential Revision: https://phabricator.services.mozilla.com/D47584
--HG--
extra : moz-landing-system : lando
Adds another method on the Emulation actor that screenshots the page content. Here, another instance of the ScreenshotActor is managed by the Emulation actor. The ScreenshotActor already provides methods to handle capturing page content along with additional options such as screenshotting the full page, which we may want to use in the future.
Differential Revision: https://phabricator.services.mozilla.com/D46092
--HG--
extra : moz-landing-system : lando
Editor listens to click events and updates location;
Outline.js does a search in symbols, adds CSS class;
Add a Mochitest
Differential Revision: https://phabricator.services.mozilla.com/D46597
--HG--
extra : moz-landing-system : lando
Depends on D46959
For those tests, using win.top was the same as using win, and did not have any impact on the asserts performed in the test.
In particular, the AutocompletePopup constructor calls for a "toolboxDoc", so passing the host document is actually more correct.
Differential Revision: https://phabricator.services.mozilla.com/D46960
--HG--
extra : moz-landing-system : lando
Depends on D46958
The getTopWindow logic was duplicated between DevToolsUtils and HTMLTooltip.
This removes the duplication and adds a fixture to the accessibility panel to avoid loading the actual DevToolsUtils file in jest tests.
Differential Revision: https://phabricator.services.mozilla.com/D46959
--HG--
extra : moz-landing-system : lando
Depends on D46957
We are updating three call sites that were still relying on window.top to get the topmost window.
For `getFrameOffsets` and `getRect`, they are currently only called with a non-null `boundaryWindow` argument.
That's why this didn't trigger any regression.
For `getFrameElement`, I could only find STRs that don't have any user impact.
For instance, when the BrowserToolbox debugs a page where you open DevTools, this will load the toolbox in an iframe.
This load will trigger `onFrameLoad` in the walker actor. And because we were using `win.top`, getFrameElement would return null, and onFrameLoad will not emit mutations.
But I couldn't see any scenario where this had an actual impact for users.
Differential Revision: https://phabricator.services.mozilla.com/D46958
--HG--
extra : moz-landing-system : lando
This changesets removes some webconsole callsites that were relying on window.top, but didn't trigger any regression.
Differential Revision: https://phabricator.services.mozilla.com/D46957
--HG--
extra : moz-landing-system : lando