"Steven Edwards" steven_ed4153@yahoo.com wrote:
Changelog: Link to WINE import libs when building programs.
I hope that this time this patch will be applied, and will not disappear without a trace as my patch "Use Wine import libraries for Programs" sent at 9 Jul 2003.
But perhaps there is a good reason to ignore it... Alexandre?
--- Dmitry Timoshkov dmitry@baikal.ru wrote:
But perhaps there is a good reason to ignore it... Alexandre?
It could have just slipped through although I can think of a good reason not to apply it now. If would be nice to have the option to use either the mingw import libs or WINEs. Our patch does hide the UUID problems that exist in w32api shell32.a.
The proper solution may be to have a configure switch like --use-native-sdk. If we could add this to configure and msvcmaker it would fix this problem and make the porting work alot simpler as I wouldnt have to manualy move headers around when build on Mingw or MS_VC.
Thanks Steven
__________________________________ Do you Yahoo!? The New Yahoo! Shopping - with improved product search http://shopping.yahoo.com
Steven Edwards steven_ed4153@yahoo.com writes:
It could have just slipped through although I can think of a good reason not to apply it now. If would be nice to have the option to use either the mingw import libs or WINEs. Our patch does hide the UUID problems that exist in w32api shell32.a.
That was my thinking, yes; the Wine programs are already tested against the Wine libraries when building on Unix, so building with a different config on Windows lets us find more problems. But if you really need the patch I can put it in, it's just that it removes some flexibility.
The proper solution may be to have a configure switch like --use-native-sdk.
With the current setup, the native libs will be used by default, and you can do "make programs LDFLAGS=-Lpath_to_wine_dlls" to use the Wine libraries. So both options are easily available.
--- Alexandre Julliard julliard@winehq.org wrote:
Steven Edwards steven_ed4153@yahoo.com writes:
It could have just slipped through although I can think of a good reason not to apply it now. If would be nice to have the option to
use
either the mingw import libs or WINEs. Our patch does hide the UUID problems that exist in w32api shell32.a.
That was my thinking, yes; the Wine programs are already tested against the Wine libraries when building on Unix, so building with a different config on Windows lets us find more problems. But if you really need the patch I can put it in, it's just that it removes some flexibility.
I would agree but for the cygwin port we really dont want to use the w32api package.
The proper solution may be to have a configure switch like --use-native-sdk.
With the current setup, the native libs will be used by default, and you can do "make programs LDFLAGS=-Lpath_to_wine_dlls" to use the Wine libraries. So both options are easily available.
My command line is already quite long when building to test the w32api headers. If you will accept a patch I will implement a option --use-mingw-sdk that does the following.
--disable-win16 CFLAGS="-D_WIN32_IE=0x0501 -D_WIN32_WINNT=0x0501 -DWINVER=0x0501 -I/mingw/include" LDFLAGS=-L/mingw/lib
Then by default on a Mingw or Cygwin build we can build all of the programs with out issue and I and anyone else can still test mingw and the w32api by using the optional switch. If you will let me attempt to implement this option then please commit the programs patch.
Thanks Steven
__________________________________ Do you Yahoo!? The New Yahoo! Shopping - with improved product search http://shopping.yahoo.com
Steven Edwards steven_ed4153@yahoo.com writes:
My command line is already quite long when building to test the w32api headers. If you will accept a patch I will implement a option --use-mingw-sdk that does the following.
--disable-win16 CFLAGS="-D_WIN32_IE=0x0501 -D_WIN32_WINNT=0x0501 -DWINVER=0x0501 -I/mingw/include" LDFLAGS=-L/mingw/lib
I don't think that should be part of a configure option. The defines should be added to the code that actually needs them, and the include and lib path cannot be hardcoded anyway.
Then by default on a Mingw or Cygwin build we can build all of the programs with out issue and I and anyone else can still test mingw and the w32api by using the optional switch. If you will let me attempt to implement this option then please commit the programs patch.
You won't be able to cleanly test other packages with the patch, since the -L$(DLLDIR) will always be present you'll most likely get a mix of Wine and Mingw libraries.