| Don't remove trailing whitespace from config file lines that are part of a list (e.g., wordlist rules). Added some const qualifiers to cfg_*() and path_*() functions. |
|
More...
|
almost 6 years ago
|
| Added const qualifiers to arguments of mem_alloc_copy() and str_alloc_copy() |
|
More...
|
almost 6 years ago
|
| 1.8.0.15: - Added linux-x86-64-avx512 and linux-x86-64-avx2 make targets, which use respectively AVX-512 and AVX2 for bitslice DES. - Dropped use of 256-bit AVX (the original "floating point" one) in 32-bit builds, dropped compile-time support for other weird bitslice DES vector types (these were for experiments only and didn't provide performance improvement on almost any real hardware). |
|
More...
|
almost 6 years ago
|
| 1.8.0.14: - Dropped undocumented limitation of the 'M' and 'Q' rule commands where they would sometimes memorize/check only up to the current hash type's length limit yet this optimization wouldn't necessarily be transparent (e.g., if a later command would extract a substring from above the hash type's length limit and bring it to within the limit). - Applied assorted other bugfixes and cleanups, some of them based on jumbo's. |
|
More...
|
almost 6 years ago
|
| Stop using the weird SSE2-16, etc. names to refer to x86-64's 16 registers extended SSE2, etc. This dates back to when x86-64 was new and exotic. |
|
More...
|
almost 6 years ago
|
| Use mem_calloc() for salt_hash |
|
More...
|
almost 6 years ago
|
| In generic build, test BF_X2 = 3 if BF_X2 = 1 was faster than BF_X2 = 0. (Change originally made in jumbo by Frank Dittrich.) |
|
More...
|
almost 6 years ago
|
| Bump RULE_RANGES_MAX from 16 to 30 |
|
More...
|
almost 6 years ago
|
| Corrected trim() not to invoke undefined behavior where it previously could set a pointer to be one char before the start of string (it wouldn't dereference that pointer, but merely setting it to point outside of the object is UB in C). |
|
More...
|
almost 6 years ago
|
| When the 'Q' command isn't last in a rule, compare the whole string (not just the portion within the current hash type's length limit, like we did before). |
|
More...
|
almost 6 years ago
|
| Re-ordered the pre-defined external mode definitions |
|
More...
|
almost 6 years ago
|
| In the 'M' command, memorize the whole string rather than just the portion needed for the 'Q' command, so that the 'X' command may extract substrings beyond the current hash type's length limit. Issue described in: https://www.openwall.com/lists/john-dev/2016/07/06/4 |
|
More...
|
almost 6 years ago
|
| Fix an off by 2 maximum length check error in the "marie -> mary" rule. This issue was reported by magnum and discussed in: https://www.openwall.com/lists/john-users/2018/12/06/1 |
|
More...
|
almost 6 years ago
|
| Set nice to 19 regardless of its previous value and hopefully portably |
|
More...
|
almost 6 years ago
|
| Call crk_done() from single_done() like other cracking modes do, from jumbo where the issue was found by Claudio and magnum (thanks!) |
|
More...
|
almost 6 years ago
|
| "rules.c bugfix: Word pair rules with implicit '+' would not update length" from jumbo, originally found & fixed by magnum (thanks!) |
|
More...
|
almost 6 years ago
|
| "Fix an academic memory leak (only happens at exit) in single.c" from jumbo, originally found by JimF and fixed by magnum (thanks!) |
|
More...
|
almost 6 years ago
|
| Added some const qualifiers |
|
More...
|
almost 6 years ago
|
| Fixed operator precedence in the external mode compiler to be the same as in C |
|
More...
|
almost 7 years ago
|
| 1.8.0.12: Dropped use of our custom int64 in favor of standard uint64_t |
|
More...
|
about 7 years ago
|
| Uncommented the linux-ppc64-altivec make target listing, added -fno-strict-aliasing to OPT_INLINE for linux-ppc*-altivec targets. |
|
More...
|
about 7 years ago
|
| Corrected the vst() macro to build as linux-ppc64-altivec on GCC Compile Farm's gcc110 (CentOS 7 default gcc). linux-ppc32-altivec also still works. |
|
More...
|
about 7 years ago
|
| MD5_IMM 1 turns out to result in much smaller and slightly faster code at least in OpenMP builds on GCC Compile Farm's gcc110 (CentOS 7, default gcc, POWER7, 64 logical CPUs). |
|
More...
|
about 7 years ago
|
| Completed dropping of the old smallest bitmap size for non-bitslice DES as well |
|
More...
|
about 7 years ago
|
| Dropped the smallest bitmap size (4-bit index, which was normally unused anyway) and introduced a new largest one (30-bit index). |
|
More...
|
about 7 years ago
|
| 1.8.0.11: - DokuWiki" external mode sample has been added to the default john.conf. - Increased SALT_HASH_LOG from 12 to 20, just like it had been done in jumbo. This should be affordable, even though it is wasteful for some formats (some have smaller salts, but currently have no way to declare that). - Dropped the smallest bitmap size (4-bit index, which was normally unused anyway) and introduced a new largest one (30-bit index). |
|
More...
|
about 7 years ago
|
| Dropped the smallest bitmap size (4-bit index, which was normally unused anyway) and introduced a new largest one (30-bit index). |
|
More...
|
about 7 years ago
|
| Increased SALT_HASH_LOG from 12 to 20, just like it had been done in jumbo. This should be affordable, even though it is wasteful for some formats (some have smaller salts, but currently have no way to declare that). |
|
More...
|
about 7 years ago
|
| Introduced PH_MASK_* just like it had been done in jumbo, to simplify merging of further changes into jumbo. |
|
More...
|
about 7 years ago
|
| "DokuWiki" external mode sample has been added to the default john.conf |
|
More...
|
about 7 years ago
|
