Open Hub
Open Hub

 Settings |  Report Duplicate

22
I Use This!
Activity Not Available

Commits : Listings

Analyzed about 1 year ago. based on code collected about 1 year ago.
Jan 18, 2023 — Jan 18, 2024
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
In the 'X' command, handle zero-length memorized strings correctly: avoid undesired integer promotion to let the 'm' variable value (last character position) overflow from 255 back to 0 (length). Thanks to magnum and JimF: http://www.openwall.com/lists/john-dev/2015/12/03/1
Solar Designer
as solar
More... about 9 years ago
Document set_key() as potentially over-reading for up to PLAINTEXT_BUFFER_SIZE total, and provide fmt_null_key for such use.
Solar Designer
as solar
More... over 9 years ago
Added atoi16l[] ("l" standing for lowercase only) and made use of it in AFS and dummy formats.
Solar Designer
as solar
More... over 9 years ago
Clarified what value the "v" command sets the "l" variable to.
Solar Designer
as solar
More... over 9 years ago
Clarified that "\p" refers to all preceding ranges, which is what makes it different from "\p0".
Solar Designer
as solar
More... over 9 years ago
Characters 2 and 3 of tripcode passwords form a descrypt salt. Formally, 8-bit characters are invalid in descrypt salts, but our implementation, as well as most others, actually handles them in some way, and the 8th bit is not necessarily ignored there. Hence, we set FMT_8_BIT, even though the 8th bit is ignored for most character positions and its behavior for positions 2 and 3 is not precisely defined. This was found by Kai Zhao's testing: http://www.openwall.com/lists/john-dev/2015/08/14/58
Solar Designer
as solar
More... over 9 years ago
Kai Zhao's extended JtR self-test in jumbo found that the bsdicrypt format wasn't always ignoring the 8th bit in characters contrary to it not having the FMT_8_BIT flag set. This commit corrects that, matching the behavior of (at least) the FreeSec implementation, and adds a test vector that would have detected the problem.
Solar Designer
as solar
More... over 9 years ago
Dropped Ultrix and SCO support
Solar Designer
as solar
More... over 9 years ago
Dropped SCO support (which probably was broken with other changes over the years anyway).
Solar Designer
as solar
More... over 9 years ago
Dropped Ultrix support (which probably was broken with other changes over the years anyway).
Solar Designer
as solar
More... over 9 years ago
In x86-64.S: When computing LM hashes, cache B(4) and B(36) in a register between S8 and S1. To completely avoid a MOV instruction, the code for S8 is modified to put its out1 in the same register where S1 expects its a6.
Solar Designer
as solar
More... over 9 years ago
1.8.0.5: - Fixed an out of bounds write bug in the external mode virtual machine. - In the generic crypt(3) format, detect descrypt with valid vs. invalid salts as separate id's for our heuristics on supported hash types. - Made a minor optimization to MMX and SSE2 assembly code (in 32-bit builds) for LM hash. - When built with --fork support, disallow session names with all-digit suffixes since these clash with those produced by --fork. - Forward SIGTERM to --fork'ed children. - Set stdout to line buffered (rather than potentially fully buffered), except for "--stdout", "--show", and auxiliary programs such as "unshadow". - On Windows, restore normal processing of Ctrl-C in case our parent (such as Johnny the GUI) had disabled it.
Solar Designer
as solar
More... over 9 years ago
Don't set stdout to line-buffered when running with --stdout or --show since these may produce a lot of output.
Solar Designer
as solar
More... over 9 years ago
Set stdout to fully buffered for --show
Solar Designer
as solar
More... over 9 years ago
If we're running the main program (as opposed to e.g. unshadow), set stdout to line buffered.
Solar Designer
as solar
More... over 9 years ago
s/Win32/Windows/
Solar Designer
as solar
More... over 9 years ago
When computing LM hashes, cache B(4) and B(36) in a register between S8 and S1. To completely avoid a MOV instruction, the code for S8 is modified to put its out1 in the same register where S1 expects its a6.
Solar Designer
as solar
More... over 9 years ago
Removed the extra instruction argument from the S-box macros, since we didn't use it for improving instruction scheduling any longer
Solar Designer
as solar
More... over 9 years ago
Relaxed the recently added invalid session name check to only disallow all-digit suffixes, since these are the ones that may clash with --fork'ed child session names
Solar Designer
as solar
More... over 9 years ago
Call SetConsoleCtrlHandler(NULL, FALSE) to restore normal processing of CTRL+C input on Windows in case our parent (such as Johnny the GUI) had disabled it.
Solar Designer
as solar
More... over 9 years ago
Check for __CYGWIN__, not __CYGWIN32__. Previously, we were checking for one or both of these to distinguish between different ancient versions of Cygwin.
Solar Designer
as solar
More... over 9 years ago
Added the missing initialization of the newly added count_base64_2 variable. Thanks to magnum for noticing that this was missing.
Solar Designer
as solar
More... over 9 years ago
Disallow session names with dots since these clash with those produced by --fork. Suggested by Frank Dittrich.
Solar Designer
as solar
More... over 9 years ago
Detect descrypt with valid vs. invalid salts as separate id's for our heuristics on supported hash types. According to Frank Dittrich, recent Fedora's crypt(3) no longer supports invalid descrypt salts even when it does support descrypt.
Solar Designer
as solar
More... over 9 years ago
Avoid left shift of negative value, from Frank Dittrich:
Solar Designer
as solar
More... over 9 years ago
benchmark_format() robustness improvements based on a report by Frank Dittrich
Solar Designer
as solar
More... over 9 years ago
In GCC-specific c_execute_fast(), initially set sp to &c_stack[2] so that there's room for op_push_* to spill imm to stack even when there wasn't actually a previous top of stack value to cache (since we're at the top level). It is simpler and quicker to let them do it than to treat this as a special case in the code.
Solar Designer
as solar
More... over 9 years ago
log_flush() after writing "Waiting for ..." Suggested by Frank Dittrich.
Solar Designer
as solar
More... over 9 years ago
Forward SIGTERM to --fork'ed children, with a twist as explained in the added comments. Thanks to Mathieu Laprise for noticing that this was missing.
Solar Designer
as solar
More... over 9 years ago
Revised the explanation of "real" and "virtual" c/s rates to cover the multi-threaded case as well.
Solar Designer
as solar
More... almost 10 years ago