On Wed, 14 Dec 2005 22:42:28 +0800, you wrote:
Well, the committed patch 1) does fix a visual glitch in the app I'm working on and 2) passes all the current tests.
1) You know that doesn't make it right. I see it causing visual glitches in *three* programs here, girotel, native hh.exe and wine's built-in notepad, which does not make it wrong either 2) Indeed, you managed to break functionality that never broke before.
There are a couple of problems with your tests and fixes
It shows a problem. If you could create a test, or otherwise point to a way to reproduce your problem I would not be arguing but looking for a way to fix them all.
: it tests child windows with CS_PARENTDC style set, and it doesn't test the case when clipping rect is NULL. In my case both cases are not fulfilled and therefore we see different results.
I still feel you do not understand what you 'fixed':
- clipping is plain MSDN specified functionality. Any undocumented restrictions of the dc to a CS_PARENTDC inhered one is unlikely; - if the *special* case cliprect==NULL is not handled correctly, then the correct fix is unlikely to be not to clip in *all* cases.
Rein.