Apache OpenOffice (AOO) Bugzilla – Issue 112422
Animations broken 3.2 -> 3.2.1
Last modified: 2017-05-20 10:22:28 UTC
Open attached presentation in 3.2 and 3.2.1. 3.2.1 shows repaint problems, object appear inside of other, disappear at the end of an animation path...
Created attachment 70015 [details] broken animation
Sorry for the crap description in words. Hopping you will clearly see the difference between both version. Tested on Windows only but XP and Vista, different graphic devices and XP was on VMware -> same problem
I can reproduce this with a simpler setup: 1. Create three Rectangle shapes A, B, and C. 2. Turn B and C into a group G and place them to overlap with A. 3. Assign Fade Out animation to A, and Fade In animation to G. 4. Make the animation on G start with that on A. Start presentation and animation. At the end of the animation on G there is a short flicker and the paint error. Two factors are crucial to reproduce this: 1. There is a group involved. 2. The duration of the fade out effect on A is not shorter than the duration of the fade in effect on G. 2a. If the durations are the same then the intersecting area of A and G is missing. 2b. If the duration of the effect on A is longer then the area G-A is missing.
This looks like a clipping problem. It may have been caused by fixes for issue 105939.
Yup, that's it. I locally turned back the fix for issue 105939 and the animation is painted as expected. @thb: Can you provide a fix in the next week? Alternatively I can revert the changes so that you have more time for the fix.
@af: I think so, will have a look. thanks for the thorough analysis! @es: sure this broken in 3.2.1?! Issue 105939 went into the 3.3 code line!
Created attachment 70353 [details] The fix
@af: see attached patch for the fix. please commit that to a suitable CWS of yours, on vacation & neither bandwidth nor diskspace for hg stunts here, sadly. FWIW, this was a stupid oversight, in that the intended behaviour should have been equal to a polygonal 'OR' operation (see e.g. basegfx's solvePolygonOperationOr(), and the previous implementation), but missed the subsequent normalization steps. Which is ~indistinguishable from the correct behaviour for simple clips.
Reassigning to af for further handling. TIA!
@jacqueline: fixes your repaint bug, too.
@thb: Thanks for the fix. Applied the patch and verified that it fixes the problem.
*** Issue 112872 has been marked as a duplicate of this issue. ***
@wg: Please verify.
Verified in CWS.