Hi CreateBitmap inside gdi32.dll should not leak any handle now when you trying create a bitmap with CreateBitmap(0,0,x,x,x); it is a GetStockObject(21) we getting back with this call and that mean we do not need delete it etiher, this should take care of few gdi handle leaks we have with some program I did implemeted it in win32k now this object it is 1x1 1bpp Bitmap you getting back. we still have bug on the syscall part for NtGdiCreateBitmap if I fix thuse bugs we breaking downtotop bitmap support, it seam something miss use CreateBitmap or or NtGdiCreateBitmap or IntGdiCreateBitmap in reactos I was force remove the correct bugfix for IntGdiCreateBitmap until we figout who does send down negtive with or height that is forbien todo that in NtGdiCreateBitmap and CreateBitmap and IntGdiCreateBitmap.
----- Original Message ----- From: "James Tabor" jimtabor.rosdev@gmail.com To: "ReactOS Development List" ros-dev@reactos.org Sent: Saturday, May 10, 2008 9:13 PM Subject: Re: [ros-dev] Time has come, a call to developers
Hi! Good advice worth more than a few cents!
Wow! I'm still using wine tests for this rewrite. It's hard work and going over my notes I see the use one test for over a year. Yes that DCE thing! The way we use it hasn't changed and we still do not pass all the tests. I haven't over looked anything. I'm still working on gdi handle management and this will fix the issue but adds more questions. What is published is not correct based on what is evident points of discovery.
Thanks, James
Ref: http://svn.reactos.org/svn/reactos/trunk/reactos/include/reactos/win32k/ntgd...
On Fri, May 9, 2008 at 6:59 PM, Alex Ionescu ionucu@videotron.ca wrote:
My two cents:
You guys shouldn't even release 0.3.5 until *all* Winetests give the same results as testing on WinXP/2003.
I've never understood why such critical/useful tests have always been ignored -- I remember fixing a lot of the ntdll regressions (thousands of them) and it took me less than a week. Mostly looking at Wine code and looking for bugs in functions that were being tested.
Ros-dev mailing list Ros-dev@reactos.org http://www.reactos.org/mailman/listinfo/ros-dev