CHMLIB crashes with SIGFPE on the attached file:

$ enum_chmLib crash.chm
Floating point exception


GDB says it's a modulo by zero:

Program received signal SIGFPE, Arithmetic exception.
0xf7fbcd55 in __umoddi3 () from /usr/lib/libchm.so.1
(gdb) bt
#0  0xf7fbcd55 in __umoddi3 () from /usr/lib/libchm.so.1
#1  0xf7fba035 in _chm_decompress_region (h=0x804b008, buf=0xffffd65c "\002", start=106, len=28) at chm_lib.c:1526
#2  0xf7fba336 in chm_retrieve_object (h=0x804b008, ui=0xffffd440, buf=0xffffd65c "\002", addr=0, len=28) at chm_lib.c:1596
#3  0xf7fb8b98 in chm_open (filename=0xffffda0b "crash.chm") at chm_lib.c:934
#4  0x08048842 in main (c=2, v=0xffffd854) at enum_chmLib.c:69
(gdb) up
#1  0xf7fba035 in _chm_decompress_region (h=0x804b008, buf=0xffffd65c "\002", start=106, len=28) at chm_lib.c:1526
1526        if (h->cache_block_indices[nBlock % h->cache_num_blocks] == nBlock    &&
(gdb) print h->cache_num_blocks
$1 = 0


This bug was found using American fuzzy lop:
https://packages.debian.org/experimental/afl

Disclaimer: I don't have spare CPU cycles, so I fuzzed only till the
first crash (which took a few seconds). It's likely that extensive
fuzzing would uncover more interesting crashers. I'd encourage CHMLIB
maintainers to perform fuzzing with AFL on their own. :-)