FS#2738 - gcc 4.0 result in weird...

Attached to Project: Arch Linux
Opened by AqD (aquila_deus) - Tuesday, 17 May 2005, 17:03 GMT
Task Type Bug Report
Category Packages: Testing
Status Closed
Assigned To No-one
Architecture not specified
Severity Critical
Priority Normal
Reported Version 0.7 Wombat
Due in Version Undecided
Due Date Undecided
Percent Complete 0%
Votes 0
Private No

Details

gcc 4.0.0-1

compile gcin-0.9.5/gcin2tab.c

LINE 165:

char fname_cin[64];
char fname_tab[64];
char tt[512];
u_char *cmd, *arg;
struct TableHead th;
struct TableHead2 th2;

....

LINE 210:

bzero(&th,sizeof(th));


the bzero() in line 210 unexpectedly clears the contents of "fname_tab".

compiled with -O2 -fomit-frame-pointer -march=i686, gcc gives no error.


with gcc-3.4.3 everything is ok.
This task depends upon

Closed by  Judd Vinet (judd)
Tuesday, 26 July 2005, 17:25 GMT
Reason for closing:  Fixed
Additional comments about closing:  fixed in gcc 4.0.1
Comment by AqD (aquila_deus) - Thursday, 19 May 2005, 14:16 GMT
PS: with glibc-2.3.5-4 and binutils-2.15.94.0.2.2-1
Comment by Indan Zupancic (i3839) - Friday, 20 May 2005, 17:40 GMT
Post this to the gcc bugtracker instead, Arch can do nothing about this. a A work around is replacing the bzero with memset, if that works.
Comment by Jan de Groot (JGC) - Friday, 20 May 2005, 21:34 GMT
Updated gcc to the latest public released snapshot, see if this fixes your problem.
Comment by AqD (aquila_deus) - Saturday, 21 May 2005, 12:48 GMT
Oh, I don't mean to tell arch guys to fix the bug. But the version of gcc should be removed, since the bug is very critical and cannot be detected at compiling time. An user may unintentionally use it to compile the kernel....
Comment by Jan de Groot (JGC) - Saturday, 21 May 2005, 14:27 GMT
gcc 4.0.0 is a bit broken and creates bad code, the new snapshot should fix those issues. KDE didn't blacklist 4.0.0 for nothing. Does the new 20050514 snapshot fix your problem?
Comment by AqD (aquila_deus) - Saturday, 28 May 2005, 12:47 GMT
I tried gcc-4.0.0-2 - gcin2tab segment faults when converting some tables during "make".
Comment by Judd Vinet (judd) - Monday, 25 July 2005, 17:53 GMT
How does gcc 4.0.1 fare?
Comment by arjan timmerman (blaasvis) - Tuesday, 26 July 2005, 15:28 GMT
just builded gcin 1.0 with gcc 4.0.1 perfectly need no extra work at all

Loading...