I'm afraid you are correct. My method of identifying stubs is broken. :( Basically I just select where version = wine and api = $api. If no rows are returned it is unimplemented. That part seems fine. Problem is when I check status to see if it is "stub". It seems there are multiple entries in the wine spec files. BitBlt is listed as a stub in 'display' but it's also cdecl in x11drv, pascal16 is gdi, and stdcall in gdi32. Perhaps I should be limiting my query to the dll which the app is trying to import from. Would that always be correct?

I agree completely. Non-windows dlls don't belong in the output since those API are not provided by windows and will never be implemented in wine. I think it's best if they're removed from the database completely rather than having the script ignore them but I'm afraid when many apps are added it could become a maintenance nightmare.

Vincent Béron wrote:

Le lun 04/08/2003 à 10:39, Dimitrie O. Paun a écrit :


On August 3, 2003 03:38 am, David Miller wrote:


Recently I spent some time on this again and would like anyone interested
in the database to take a look and email me comments.


When comparing 'exports' for example, I get this kind of output:
...
IID_IAVIStream is exported by win98-4.10.1998 and wine
IID_IGetFrame is exported by win98-4.10.1998 and wine
BitReverseBuf is exported by win98-4.10.1998 only
FaxCodecChange is exported by win98-4.10.1998 only
...

which is fine, but I will drop the similarities:

IID_IAVIStream is exported by win98-4.10.1998 and wine
IID_IGetFrame is exported by win98-4.10.1998 and wine

just output the differences, there are too many anyway.
Or have a checkbox to ask the user if they want to see
only the differences.



I checked the unimplemented APIs for vcdeasy, and there are two things: - Cygwin's dlls (cygz, cygjpeg6b, cygxml2-2, cygwin1) are listed, when they're not part of a Windows installation. You should try to separate what's part of Windows (any version), and what can be found elsewhere (from an application, a toolkit, etc.). - Also, in gdi32, BitBlt and StretchBlt are marked as stubs. I have a hard time believing that. What's your method to identify stubs?

Vincent












Reply via email to