Skip Menu |

This queue is for tickets about the Imager CPAN distribution.

Report information
The Basics
Id: 1388
Status: resolved
Priority: 0/
Queue: Imager

People
Owner: TONYC [...] cpan.org
Requestors: addi [...] umich.edu
Cc:
AdminCc:

Bug Information
Severity: (no value)
Broken in: (no value)
Fixed in: (no value)



Subject: Imager-0.41 5.8.0 test segfault in tga test
Hi, the t/t108tga.t test generates a segfault for me using 5.8.0. It works fine under 5.6.1 so I'm not sure what you're tickling. Here's a stack trace, output of perl-v attached: Starting program: /usr/local/perl5.8.0-g/bin/perl -Mblib t/t108tga.t 1..16 ok 1 ok 2 ok 3 ok 4 not ok 5 # not ok 6 # not ok 7 # not ok 8 # not ok 9 # not ok 10 # ok 11 not ok 12 # Targa Image has none of 15/16/24/32 pixel layout: could not read targa colourmap not ok 13 # Targa Image has none of 15/16/24/32 pixel layout: could not read targa colourmap not ok 14 # could not read targa colourmap: could not read targa colourmap Program received signal SIGSEGV, Segmentation fault. 0x0fe36e18 in free () from /lib/libc.so.6 (gdb) bt #0 0x0fe36e18 in free () from /lib/libc.so.6 #1 0x0fe3766c in realloc () from /lib/libc.so.6 #2 0x0fe37104 in realloc () from /lib/libc.so.6 #3 0x1006c3c0 in Perl_safesysrealloc (where=0x1035fdf0, size=26) at util.c:122 #4 0x10089f10 in Perl_sv_grow (sv=0x1017c8bc, newlen=26) at sv.c:1580 #5 0x1009a7fc in Perl_sv_catpvn_flags (dsv=0x1017c8bc, sstr=0x1019bf10 "3553974", slen=7, flags=271997336) at sv.c:4270 #6 0x1007fcb0 in Perl_pp_concat () at pp_hot.c:187 #7 0x1006b624 in Perl_runops_debug () at dump.c:1398 #8 0x100176ec in S_run_body (oldscope=1) at perl.c:1681 #9 0x100172a0 in perl_run (my_perl=0xfeedcc0) at perl.c:1600 #10 0x10014174 in main (argc=3, argv=0x7ffff884, env=0x10365998) at perlmain.c:85 #11 0x0fddac30 in __libc_start_main () from /lib/libc.so.6
I couldn't reproduce this with current CVS. The original reporter could try to set IM_DEBUG_MALLOC and see if that catches anything. Given the stack trace it could be anything from an out of bounds write to a double-free, in either perl or Imager. IM_DEBUG_MALLOC would give us some idea which. Tony
[guest - Sun Aug 4 02:32:51 2002]: Show quoted text
> Hi, the t/t108tga.t test generates a segfault for me using 5.8.0. It > works fine under 5.6.1 so I'm not sure what you're tickling.
I can't reproduce this with 5.8.0 built under Debian 3.0 with Imager 0.41 either. The original report mentioned perl-v (hopefully -V) information, but didn't include it. Is it available?
given the failures in the write tests (and the types of failures in the read tests) I wonder if the user ran out of disk space.
Could not reproduce.