more valgrind complaints

[Matthias Andree]

Hi,

I have "interesting"  output during bogoutil runs, I'm starting to
suspect DB-4.0.14 memory allocation issues. I have more of them, same
place, and I wonder why DB allocs only 2 or 3 bytes for the key data.

bogoutil.c:71 is the c_get line, bogoutil.c:90 is the printf that prints
the key and the count (no date).

Can anyone run their version against valgrind? I'm committing a patch to
make this easy.

==18661== Invalid read of size 1
==18661==    at 0x4035C935: __strnlen (in /lib/libc.so.6)
==18661==    by 0x4033199A: _IO_vfprintf (in /lib/libc.so.6)
==18661==    by 0x40338CC1: _IO_printf (in /lib/libc.so.6)
==18661==    by 0x8048E92: dump_file (bogoutil.c:90)
==18661==    by 0x403014A1: __libc_start_main (in /lib/libc.so.6)
==18661==    by 0x8048C00: (within /home/emma/cvs-3rdparty/bogofilter-mime/bogoutil)
==18661==    Address 0x410F359E is 0 bytes after a block of size 2 alloc'd
==18661==    at 0x4015D50F: malloc (vg_clientfuncs.c:100)
==18661==    by 0x402C43F7: __os_malloc (in /usr/lib/libdb-4.0.so)
==18661==    by 0x402C44C5: __os_realloc (in /usr/lib/libdb-4.0.so)
==18661==    by 0x40292131: __db_retcopy (in /usr/lib/libdb-4.0.so)
==18661==    by 0x40291FA4: __db_ret (in /usr/lib/libdb-4.0.so)
==18661==    by 0x40283630: __db_c_get (in /usr/lib/libdb-4.0.so)
==18661==    by 0x8048D84: dump_file (bogoutil.c:71)
==18661==    by 0x403014A1: __libc_start_main (in /lib/libc.so.6)
==18661==    by 0x8048C00: (within /home/emma/cvs-3rdparty/bogofilter-mime/bogoutil)

-- 
Matthias Andree