Follow-up Comment #2, patch #4415 (project gnustep):
I think it would be safer to tell the world that win32 works differently in
the header file. For example, the change to filesystemRepresentation broke
my code, and I didn't even know it. I updated by gnustep-base library to a
new version, and didn't get a compiler warning or anything, but
filesystemRepresentation only returned the first character of the string.
(From my perspective interpreting it as a char*)
The current functionally is flat out dangerous. Either return a char* or
change the header to what you will actually return. You can't expect every
user of -[NSString filesystemRepresentation] to read the source code of
gstep-base.
_______________________________________________________
Reply to this item at:
<http://savannah.gnu.org/patch/?func=detailitem&item_id=4415>
_______________________________________________
Message sent via/by Savannah
http://savannah.gnu.org/
_______________________________________________
Bug-gnustep mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/bug-gnustep
