I just committed a patch that completely removes the win32.cpp/.h that were causing build problems (and apparently weren't used anymore). Apart from ripping out some printing related code that wasn't used anymore this also removed the is_os_wide function. Although this function is in principle unrelated and could have been kept, I decided to get rid of that as well. This means that any remaining support for Win9x/ME that depended on this is now gone (as far as know we didn't support those windows versions anymore anyway, so it was just extra baggage).
So, what next? I might have encountered a pretty serious regression in text rendering, but I haven't looked deeper into that yet. (And it might be known already or it might be something specific to my set up.)
2011/4/15 Jasper van de Gronde <th.v.d.gronde@...528...>:
I just committed a patch that completely removes the win32.cpp/.h that were causing build problems (and apparently weren't used anymore). Apart from ripping out some printing related code that wasn't used anymore this also removed the is_os_wide function. Although this function is in principle unrelated and could have been kept, I decided to get rid of that as well. This means that any remaining support for Win9x/ME that depended on this is now gone (as far as know we didn't support those windows versions anymore anyway, so it was just extra baggage).
We are depending on versions of Glib and GTK which don't work on pre-NT Windows, so this support was meaningless anyway.
So, what next? I might have encountered a pretty serious regression in text rendering, but I haven't looked deeper into that yet. (And it might be known already or it might be something specific to my set up.)
Besides fixing the text bug we need to fix the rendering of controls. There might also be some unfixed rendering problems. I remember that zero width / zero height rectangles are rendered differently in the trunk than in Cairo branch.
Regards, Krzysztof
participants (2)
-
Jasper van de Gronde -
Krzysztof Kosiński