GNOME Bugzilla – Bug 562196
build error with gcc 4.3.2
Last modified: 2008-11-25 23:38:22 UTC
building 2.25.8 with gcc 4.3.2: cc1: warnings being treated as errors core/xprops.c: In function 'wm_hints_from_results': core/xprops.c:729: error: signed and unsigned type in conditional expression core/xprops.c:732: error: signed and unsigned type in conditional expression core/xprops.c:733: error: signed and unsigned type in conditional expression core/xprops.c: In function 'size_hints_from_results': core/xprops.c:853: error: signed and unsigned type in conditional expression core/xprops.c:854: error: signed and unsigned type in conditional expression core/xprops.c:855: error: signed and unsigned type in conditional expression core/xprops.c:856: error: signed and unsigned type in conditional expression core/xprops.c:857: error: signed and unsigned type in conditional expression core/xprops.c:858: error: signed and unsigned type in conditional expression core/xprops.c:859: error: signed and unsigned type in conditional expression core/xprops.c:860: error: signed and unsigned type in conditional expression core/xprops.c:861: error: signed and unsigned type in conditional expression core/xprops.c:862: error: signed and unsigned type in conditional expression core/xprops.c:863: error: signed and unsigned type in conditional expression core/xprops.c:864: error: signed and unsigned type in conditional expression core/xprops.c:865: error: signed and unsigned type in conditional expression core/xprops.c:866: error: signed and unsigned type in conditional expression core/xprops.c:871: error: signed and unsigned type in conditional expression core/xprops.c:872: error: signed and unsigned type in conditional expression core/xprops.c:873: error: signed and unsigned type in conditional expression
Created attachment 123356 [details] [review] Cast to unsigned int Can you tell me whether this patch solves the problem for you?
It probably would, but I've already nuked the evil -Werror from CFLAGS. Of course, it makes those macro names horribly misleading...
(In reply to comment #2) > It probably would, but I've already nuked the evil -Werror from CFLAGS. It's not much help for other people, though (unless we were to turn -Werror off in svn). > Of course, it makes those macro names horribly misleading... Not really. What's misleading is people passing in signed values to a macro which expects unsigned values. Or would you prefer the cast to be added to every relevant invocation of the macro?
*** Bug 562106 has been marked as a duplicate of this bug. ***
This problem has been fixed in the development version. The fix will be available in the next major software release. Thank you for your bug report.