I've found a bug in vim, i have no time for debugging, but here's an example and a minor backtrace:
looks like %edx is used as a pointer and this pointer is NULL. No idea if this should be fixed in this way, but here's some info: [EMAIL PROTECTED] vim -r /tmp/ Vim: Caught deadly signal SEGV Press ENTER or type command to continueVim: Finished. Segmentation fault 0x080e86d4 in ml_recover () (gdb) bt #0 0x080e86d4 in ml_recover () #1 0x080ddc38 in main () 0x080e86cc <ml_recover+284>: mov $0x1,%bl 0x080e86ce <ml_recover+286>: mov 0xfffffeec(%ebp),%edx 0x080e86d4 <ml_recover+292>: mov 0xc(%edx),%eax <---------here 0x080e86d7 <ml_recover+295>: mov %eax,(%esp) 0x080e86da <ml_recover+298>: call 0x80fe440 <vim_free> 0x080e86df <ml_recover+303>: mov 0xfffffeec(%ebp),%ecx 0x080e86e5 <ml_recover+309>: mov %ecx,(%esp) 0x080e86e8 <ml_recover+312>: call 0x80fe440 <vim_free> (gdb) info registers eax 0x0 0 ecx 0x81b7fc0 136019904 edx 0x0 0 ebx 0x1 1 esp 0xbf85df10 0xbf85df10 ebp 0xbf85e048 0xbf85e048 esi 0x0 0 edi 0x3 3 eip 0x80e86d4 0x80e86d4 <ml_recover+292> eflags 0x10246 [ PF ZF IF RF ] cs 0x73 115 ss 0x7b 123 ds 0x7b 123 es 0x7b 123 fs 0x0 0 gs 0x0 0 (gdb) --pancake