https://bugs.documentfoundation.org/show_bug.cgi?id=159213

--- Comment #12 from Michael Weghorn <[email protected]> ---
Thanks for the fix, Julien!

(In reply to Robert Großkopf from comment #0)
> Bug doesn't appear with LO 7.4.7.2 (my working version for Base - too many
> bugs in LO 7.5 and 7.6 to create databases there). So it's a regression.

This most likely started crashing with the following commit that enabled
handling for the corresponding a11y CHILD event, i.e. the underlying issue was
there before, but only started to cause a crash once more accessibility
handling was implemented for the Qt-based VCL plugins.

commit be8b031d8b3c66b223ea2478f1129427f3a362bd
Author: Michael Weghorn
Date:   Thu Sep 7 09:51:38 2023 +0200

    qt a11y: Forward CHILD event for removed child

    No longer comment out the code to send a
    `QAccessible::ObjectDestroyed` event when
    receiving an `AccessibleEventId::CHILD` event with
    its `OldValue` set.

    The underlying issues causing crashes previously seem
    to be fixed, I can no longer reproduce these on Debian
    testing with Orca 44.1-2 when using the font color
    popup as described in

        commit 734d10ed3612d75edcee145475ddd0b0165efeac
        Author: Michael Weghorn <[email protected]>
        Date:   Fri Apr 14 16:57:09 2023 +0300

            qt a11y: Send QAccessible::ObjectCreated event for correct object

    > Adapting this for the case where a child has
    > been removed (bridged to Qt as `QAccessible::ObjectDestroyed`
    > event) would currently results in crashes when closing the
    > application e.g. after using the character font color popup in
    > the Writer toolbar. This needs further investigation, so don't
    > send the event for now, but add a `SAL_WARN`.

    Change-Id: Ib8f21850dd56645cf64a74be0e1ff8242615b928
    Reviewed-on: https://gerrit.libreoffice.org/c/core/+/156647
    Tested-by: Jenkins
    Reviewed-by: Michael Weghorn

(In reply to Michael Weghorn from comment #8)
> And a possible consideration might then be to revise + fix all of the places
> where a CHILD event is sent, and drop the workaround (just on master) again,
> once all of these should be safe. (I'm happy to do the latter as I find time
> if it makes sense.)

I've added that to my task list and will come back to it as I find time.

-- 
You are receiving this mail because:
You are the assignee for the bug.

Reply via email to