John the Ripper jumbo - advanced offline password cracker, which supports hundreds of hash and cipher types, and runs on many operating systems, CPUs, GPUs, and even some FPGAs
I have found a bug that may be more of a 'class of bug' in JtR.
_Specifically, LMfmt defines 'split' to return a static buffer. At at least one point, that static buffer is returned into a variable, and then split is later called, and the result stored into a second variable. It is
unfortunately expected to compare these two variables, but since they are a static buffer, the comparison is always the same.
_Specifically, this affects crk_reload_pot() (which calls split) and crk_remove_potentry() (which also calls split, via LM's 'source').
_This was affecting my cracking because on windows anyway, crk_reloadpot gets called by the simulated timer during normal operation on a regular basis. As soon as that happened, it would actually stop cracking because the 'remove hash' stuff would always remove all the remaining hashes.
I also got a patch that look solid. I will commit that very soon but first I'll try to reproduce - I can't believe this never showed up in my very extensive testing.
Got report OOB (and via Solar).
I have found a bug that may be more of a 'class of bug' in JtR.
_Specifically, LMfmt defines 'split' to return a static buffer. At at least one point, that static buffer is returned into a variable, and then split is later called, and the result stored into a second variable. It is unfortunately expected to compare these two variables, but since they are a static buffer, the comparison is always the same.
_Specifically, this affects crk_reload_pot() (which calls split) and crk_remove_potentry() (which also calls split, via LM's 'source').
_This was affecting my cracking because on windows anyway, crk_reloadpot gets called by the simulated timer during normal operation on a regular basis. As soon as that happened, it would actually stop cracking because the 'remove hash' stuff would always remove all the remaining hashes.
I also got a patch that look solid. I will commit that very soon but first I'll try to reproduce - I can't believe this never showed up in my very extensive testing.