https://bugs.freedesktop.org/show_bug.cgi?id=60910

Miklos Vajna <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[email protected]

--- Comment #11 from Miklos Vajna <[email protected]> ---
So it turns out the 4-0 branch-off was still OK; however latest 4-0 doesn't
export the shape. Bisect says:

75118d2f9ebdf96250d3a4c78b695085c3527e4e is the first bad commit
commit 75118d2f9ebdf96250d3a4c78b695085c3527e4e
Author: David Tardon <[email protected]>
Date:   Tue Dec 18 15:24:28 2012 +0100

    fdo#56267, fdo#56980 propagate shape change to subclasses

    It turns out (as witnessed by fdo#56267) that my fix for fdo#56980 only
    cured the symptom, not the cause. The real problem is caused by the
    following sequence of events during ODF import:

    1) an SvxCustomShape object is created (XShape iface)
    2) an SdrObjCustomShape object is created for the SvxCustomShape, but it
       is not associated with it (yet)
    3) another SvxCustomShape object is created internally by the
       SdrObjCustomShape and they are associated
    4) an EnhancedCustomShapeEngine is created for this SvxCustomShape by
       SdrObjCustomShape
    5) the SvxCustomShape from point 1 is set to the SdrObjCustomShape

    At some point (I did not follow this explicitly) the SvxCustomShape
    cached by the EnhancedCustomShapeEngine loses its (weak) reference to
    the SdrObjCustomShape. This leaves it gutted and all subsequent calls to
    render() return an empty XShape.

    The solution is simple: let SdrObjCustomShape know that the associated
    UNO shape has changed, so it can drop the custom shape engine.

    Change-Id: I267838ea4857dfcd646f40c811f3ae572237a1e6
    (cherry picked from commit 7fec8dfcaca4efc92516f9af51a3157f1a11ccd7)

David, sorry to hear, it seems we have one more commit in this never-ending
game. ;-(

-- 
You are receiving this mail because:
You are the assignee for the bug.
_______________________________________________
Libreoffice-bugs mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/libreoffice-bugs

Reply via email to