Eric Blake wrote:
> I've had this sitting around for a while [1], so now that we have
> confirmed this is a problem in a wild, I'm committing it.
Very nice! I'm adding a mention of the ia64 specific bug that Simon
encountered.
2009-06-14 Bruno Haible <br...@clisp.org>
* m4/memchr.m4: Mention also the bug on IA-64.
* doc/posix-functions/memchr.texi: Likewise.
--- doc/posix-functions/memchr.texi.orig 2009-06-14 23:38:55.000000000
+0200
+++ doc/posix-functions/memchr.texi 2009-06-14 23:26:21.000000000 +0200
@@ -13,7 +13,7 @@
@item
This function dereferences too much memory on some platforms:
-glibc 2.10 on x86_64 or Alpha
+glibc 2.10 on x86_64, IA-64, Alpha.
@end itemize
Portability problems not fixed by Gnulib:
--- m4/memchr.m4.orig 2009-06-14 23:38:55.000000000 +0200
+++ m4/memchr.m4 2009-06-14 23:31:21.000000000 +0200
@@ -25,6 +25,7 @@
# http://bugzilla.redhat.com/499689
# memchr should not dereference overestimated length after a match
# http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=521737
+ # http://sourceware.org/bugzilla/show_bug.cgi?id=10162
# Assume that memchr works on platforms that lack mprotect.
AC_CACHE_CHECK([whether memchr works], [gl_cv_func_memchr_works],
[AC_RUN_IFELSE([AC_LANG_PROGRAM([[
