Found in

• Firefox 111.0b8

Affected versions

• Firefox 111.0b8
• Firefox 112.0a1

Tested platforms

• Affected platforms: macOS 12/13
• Unaffected platforms: Windows, Ubuntu

Steps to reproduce

1. Launch Firefox
2. Access an editable pdf, for example
3. Select the Draw pen and draw anything on the pdf
4. Right click over the pdf and close the context menu after

Expected result

• The drawing pen is still visible

Actual result

• The default cursor icon is used

Regression range

• Will see for a regression

Additional notes

• The issue can be seen in the following attachment
• Windows and Ubuntu keeps the draw pen cursor after closing the context menu.

Seems that this is not a regression as it happens on Firefox 104.0a1 (2022-07-19), the day the PDF drawing functionality was added.

I can reproduce either the issue with the html file I attached so it isn't a pdf.js issue.
:spohl, maybe you'd have an idea on what's wrong here.

This appears to have stopped working with native context menus. Setting widget.macos.native-context-menus to false doesn't reproduce the issue.

:mstange, since you are the author of the regressor, bug 1698997, could you take a look?

For more information, please visit auto_nag documentation.

These guards are only needed at the boundary to C++, see https://bugzilla.mozilla.org/show_bug.cgi?id=1693392 .

Happy code deletion day!

nsMacCursor was used so that animated cursors could be treated like non-animated cursors.
But we don't have animated cursors anymore! Nothing calls -[nsMacCursor cursorWithFrames:type:].

Depends on D172479

Depends on D172480

Depends on D172481

Depends on D172482

This renames nsMacCursorManager to MOZDynamicCursor and makes it implement NSCursor.

We also install a cursor rect which covers the entire ChildView and which refers
to the MOZDynamicCursor singleton.

With the help of the cursor rect, the -[NSCursor set] implementation is now called
at the right times.

We can also stop resetting the cursor when it leaves the window.

Depends on D172483

Set release status flags based on info from the regressing bug 1698997

Backed out 6 changesets (Bug 1820168) for causing multiple failures CLOSED TREE

Push with failures: https://treeherder.mozilla.org/jobs?repo=autoland&group_state=expanded&collapsedPushes=842382&selectedTaskRun=AJSZBIsGQd2FFDAfAz-g6g.0&resultStatus=testfailed%2Cbusted%2Cexception%2Cretry%2Cusercancel&revision=03bc416d339956f046f032ed688cc28e19555505

Log: https://treeherder.mozilla.org/logviewer?job_id=408884863&repo=autoland&lineNumber=4302
https://treeherder.mozilla.org/logviewer?job_id=408884902&repo=autoland&lineNumber=5687

Backout: https://hg.mozilla.org/integration/autoland/rev/bd439d90e560048f09bc7424127df84caff68a6c

There are some r+ patches which didn't land and no activity in this bug for 2 weeks.
:mstange, could you have a look please?
If you still have some work to do, you can add an action "Plan Changes" in Phabricator.
For more information, please visit auto_nag documentation.

(In reply to Release mgmt bot [:suhaib / :marco/ :calixte] from comment #15)

There are some r+ patches which didn't land and no activity in this bug for 2 weeks.

More work is needed to find the reason for the crashes that were the reason for the backout.
This has exceeded my time box for non-speedometer work at the moment, so I'll make this bug available for anyone who wants to finish it.