Marcus Crestani <marcus(a)crestani.de&gt; wrote: Wow, you're fast! I just checked that stuff in a few hours ago. Obviously, I didn't test it out on a sufficiently large number of gcc versions, though. Here's a fix, with the added bonus that I get to check in the regeneration of depend that I forgot to include with the last patch. APPROVE COMMIT 21.5 src/ChangeLog addition: 2003-10-14 Jerry James <james(a)xemacs.org&gt; * compiler.h: Check __GNUC_PATCHLEVEL__ definition. * depend: Regenerate. xemacs-21.5 source patch: Diff command: cvs -q diff -uN Files affected: src/depend src/compiler.h Index: src/compiler.h =================================================================== RCS file: /pack/xemacscvs/XEmacs/xemacs/src/compiler.h,v retrieving revision 1.1 diff -d -u -u -r1.1 compiler.h --- src/compiler.h 2003/10/14 05:03:03 1.1 +++ src/compiler.h 2003/10/14 14:33:45 ＠＠ -75,12 +75,17 ＠＠ # endif #endif -/* Macro simplification for non-GNU compilers */ +/* Macro simplification for non-GNU compilers and older gccs that did not + define all of these symbols */ #ifndef __GNUC__ #define __GNUC__ 0 +#endif +#ifndef __GNUC_MINOR__ #define __GNUC_MINOR__ 0 +#endif +#ifndef __GNUC_PATCHLEVEL__ #define __GNUC_PATCHLEVEL__ 0 -#endif /* __GNUC__ */ +#endif /* Simplify testing for specific GCC versions. This also works for non-GCC compilers, where GCC_VERSION is zero. */ Index: src/depend =================================================================== RCS file: /pack/xemacscvs/XEmacs/xemacs/src/depend,v retrieving revision 1.42 diff -d -u -u -r1.42 depend --- src/depend 2003/09/16 03:57:58 1.42 +++ src/depend 2003/10/14 14:33:45 ＠＠ -11,7 +11,7 ＠＠ LISP_H= #else CONFIG_H=config.h -LISP_H=lisp.h config.h dumper.h general-slots.h lrecord.h symeval.h symsinit.h text.h $(LISP_UNION_H) +LISP_H=lisp.h compiler.h config.h dumper.h general-slots.h lrecord.h symeval.h symsinit.h text.h $(LISP_UNION_H) #endif #if defined(HAVE_MS_WINDOWS) -- Jerry James http://www.ittc.ku.edu/~james/