|
Re: [patch]: fix ia64 SIGSEGV in coreutils 4.5.9: msg#00046gnu.core-utils.bugs
David Kimdon <dwhedon@xxxxxxxxxx> wrote: > The included patch fixes the Debian ia64 build failure for coreutils-4.5.9 > reported here: > > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=183962 Thanks a lot for the nice report and fix. I've fixed that slightly differently for 4.5.10. 2003-03-07 Jim Meyering <jim@xxxxxxxxxxxx> * getcwd.c: #undef getcwd only after *all* included header files. Declare getcwd unconditionally. The problem was that same.h included config.h, and that defined getcwd to rpl_getcwd a second time (after the original #undef), and led to the rpl_getcwd function calling itself endlessly. This was the cause of at least the `du' test failures on IRIX 6.5, reported by Nelson Beebe. Index: lib/getcwd.c =================================================================== RCS file: /fetish/cu/lib/getcwd.c,v retrieving revision 1.2 retrieving revision 1.5 diff -u -p -u -p -r1.2 -r1.5 --- lib/getcwd.c 5 Mar 2003 08:19:10 -0000 1.2 +++ lib/getcwd.c 7 Mar 2003 10:13:55 -0000 1.5 @@ -32,16 +32,20 @@ extern int errno; #include "pathmax.h" #include "same.h" -/* Undefine getcwd here so any prototype is not redefined to be a - prototype for rpl_getcwd. */ -#undef getcwd - /* Guess high, because that makes the test below more conservative. But this is a kludge, because we should really use pathconf (".", _PC_NAME_MAX). But it's probably not worth the cost. */ #define KLUDGE_POSIX_NAME_MAX 255 #define MAX_SAFE_LEN (PATH_MAX - 1 - KLUDGE_POSIX_NAME_MAX - 1) + +/* Undefine getcwd here, as near the use as possible, in case any + of the files included above define it to rpl_getcwd. */ +#undef getcwd + +/* Any declaration of getcwd from headers included above has + been changed to a declaration of rpl_getcwd. Declare it here. */ +extern char *getcwd (char *buf, size_t size); /* This is a wrapper for getcwd. Some implementations (at least GNU libc 2.3.1 + linux-2.4.20) return |
|
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| Previous by Date: | Re: language of coreutils: 00046, Paul Eggert |
|---|---|
| Next by Date: | Re: df reports size incorrectly when used with "-B": 00046, Paul Eggert |
| Previous by Thread: | [patch]: fix ia64 SIGSEGV in coreutils 4.5.9i: 00046, David Kimdon |
| Next by Thread: | du make check failure on Solaris 5.5 and 5.8 in coreutils-4.5.9: 00046, Vin Shelton |
| Indexes: | [Date] [Thread] [Top] [All Lists] |
| News | FAQ | advertise |