RE: latest spam-milter

[Ron Snyder]

> > is SMFIS_TEMPFAIL supposed to make the program terminate and dump
> > core (sig 6) like that ?
> 
> No.  Signal 6 is just the result of someone calling abort().  
> You might
> want to run the milter without the -f flag for a while.  abort() will
> print a diagnostic to stderr before killing the program.

I've put spamass-milter into a loop, here's the output over the last 8
minutes (the timestamp is associated with the line above it):
Memory fault (core dumped)
Mon Nov 18 15:00:06 PST 2002
Memory fault (core dumped) 
Mon Nov 18 15:01:12 PST 2002
Memory fault (core dumped) 
Mon Nov 18 15:01:41 PST 2002
Memory fault (core dumped) 
Mon Nov 18 15:03:21 PST 2002
Memory fault (core dumped) 
Mon Nov 18 15:04:31 PST 2002
Memory fault (core dumped) 
Mon Nov 18 15:06:21 PST 2002
Memory fault (core dumped) 
Mon Nov 18 15:06:53 PST 2002

Here's the backtrace:
# gdb /usr/local/sbin/spamass-milter /spamass-milter.core  
GNU gdb 4.16.1
Copyright 1996 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain
conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for details.
This GDB was configured as "i386-unknown-openbsd2.7"...
Core was generated by `spamass-milter'.
Program terminated with signal 11, Segmentation fault.
Reading symbols from /usr/libexec/ld.so...done.
Reading symbols from /usr/lib/libstdc++.so.30.0...done.
Reading symbols from /usr/lib/libm.so.0.1...done.
Reading symbols from /usr/lib/libc_r.so.3.0...done.
#0  0xa6ad in SpamAssassin::read_pipe (this=0x24340) at
spamass-milter.cpp:965
965             return size;
(gdb) bt
#0  0xa6ad in SpamAssassin::read_pipe (this=0x24340) at
spamass-milter.cpp:965
#1  0xa781 in SpamAssassin::empty_and_close_pipe (this=0x24340)
    at spamass-milter.cpp:976
#2  0x9214 in SpamAssassin::input (this=0x24340) at spamass-milter.cpp:765
#3  0x5ced in mlfi_eom (ctx=0x23500) at spamass-milter.cpp:508
#4  0xf31f in st_bodyend ()
#5  0xe798 in mi_engine ()
#6  0xcad1 in mi_handle_session ()
#7  0xc221 in mi_thread_handle_wrapper ()
#8  0x40086518 in _thread_start ()
(gdb) quit

0xa680 <read_pipe__12SpamAssassin+2828>:        addl   $0x18,%esp
0xa683 <read_pipe__12SpamAssassin+2831>:        pushl  $0x9b54
0xa688 <read_pipe__12SpamAssassin+2836>:        pushl  $0x2
0xa68a <read_pipe__12SpamAssassin+2838>:        call   0xa7c8
<debug__FiPCce>
0xa68f <read_pipe__12SpamAssassin+2843>:        addl   $0x10,%esp
0xa692 <read_pipe__12SpamAssassin+2846>:        addl   $0xfffffff8,%esp
0xa695 <read_pipe__12SpamAssassin+2849>:        pushl  $0x9b62
0xa69a <read_pipe__12SpamAssassin+2854>:        pushl  $0x1
0xa69c <read_pipe__12SpamAssassin+2856>:        call   0xa7c8
<debug__FiPCce>
0xa6a1 <read_pipe__12SpamAssassin+2861>:        movl   0xfffffb18(%ebp),%eax
0xa6a7 <read_pipe__12SpamAssassin+2867>:        movl   0x4(%eax),%edx
0xa6aa <read_pipe__12SpamAssassin+2870>:        movl   0x4(%edx),%eax
0xa6ad <read_pipe__12SpamAssassin+2873>:        movl   (%eax),%eax
0xa6af <read_pipe__12SpamAssassin+2875>:        movl   %eax,0x4(%edx)
0xa6b2 <read_pipe__12SpamAssassin+2878>:        movl   0xfffffbfc(%ebp),%ecx
0xa6b8 <read_pipe__12SpamAssassin+2884>:        leal   0xfffffff0(%ecx),%edi
0xa6bb <read_pipe__12SpamAssassin+2887>:        movl   0xfffffff8(%ecx),%eax
0xa6be <read_pipe__12SpamAssassin+2890>:        leal   0xffffffff(%eax),%edx
0xa6c1 <read_pipe__12SpamAssassin+2893>:        movl   %edx,0xfffffff8(%ecx)
0xa6c4 <read_pipe__12SpamAssassin+2896>:        cmpl   $0x1,%eax

Although _most_ of them are failing on 0xa6ad, there was at least one
failure at 0xa6aa.  These backtrace and disassemble outputs are identical to
those that are generated with   MALLOC_OPTIONS="AJ" .    

-ron