| Added john_main_process, john_child_count, and john_child_pids to best.c to repair "make generic" builds, which were broken with recent commits |
|
More...
|
over 11 years ago
|
| Moved and renamed sig_pids from signals.[ch] to john_child_pids in john.[ch]. Introduced john_child_count. |
|
More...
|
over 11 years ago
|
| John's global variables exported from john.c to other source files (currently this is just john_main_process). |
|
More...
|
over 11 years ago
|
| When running --fork'ed, selectively mute some messages in child processes (those messages that are expected to be printed by the main process as well, such as on warning/error conditions that are almost certainly common to all processes). |
|
More...
|
over 11 years ago
|
| 1.7.9.12: The --fork=N option has been added |
|
More...
|
over 11 years ago
|
| Don't try to signal the (non-existent) child processes when we're not --fork'ed |
|
More...
|
over 11 years ago
|
| Prefix log file lines with node number when running a --fork'ed session |
|
More...
|
over 11 years ago
|
| Unlock the .rec file before closing it when we intend to re-open the file for writing and lock it again. This appears to help avoid a race condition on Linux where a subsequent rec_lock() would fail (flock() failing with EWOULDBLOCK) if too little time elapses yet the process happens to be bounced to a different CPU. |
|
More...
|
over 11 years ago
|
| Print the status line to stderr all at once, not piece by piece |
|
More...
|
over 11 years ago
|
| Prefix the status line with node number when running a --fork'ed session |
|
More...
|
over 11 years ago
|
| Send SIGUSR2 to child processes on a keypress, and have them print status upon receipt of this signal |
|
More...
|
over 11 years ago
|
| Implemented support for --restore and --status for --fork'ed sessions |
|
More...
|
over 11 years ago
|
| Implemented --fork |
|
More...
|
over 11 years ago
|
| Implemented --node for single crack mode |
|
More...
|
over 11 years ago
|
| Dropped FLG_NODE_SUP in favor of specifying what modes --node is incompatible with right on its line in options.c, now that it's compatible with almost all. |
|
More...
|
over 11 years ago
|
| Implemented --node for external modes |
|
More...
|
over 11 years ago
|
| Moved the rule and word variables back to do_wordlist_crack() function scope (rule could be used uninitialized otherwise) |
|
More...
|
over 11 years ago
|
| Use "unsigned long" for line numbers |
|
More...
|
over 11 years ago
|
| Use skip_lines() in restore_line_number() |
|
More...
|
over 11 years ago
|
| Moved the non-zero count check to inside skip_lines() |
|
More...
|
over 11 years ago
|
| Implemented proper interrupt/restore of sessions where we're running with a range of node numbers rather than with just one node number |
|
More...
|
over 11 years ago
|
| Corrected a design error that caused extra words to be skipped when we're running with a range of node numbers (node_max > node_min) rather than with just one node number. Proper interrupt/restore of such sessions is not implemented yet - to be committed separately. |
|
More...
|
over 11 years ago
|
| Renamed the dist* variables to dist_* for consistency with other variable names |
|
More...
|
over 11 years ago
|
| Corrected the initial skipping of other nodes' lines when restoring a session |
|
More...
|
over 11 years ago
|
| Simplified the initialization of distrules now that it's no longer a tri-state |
|
More...
|
over 11 years ago
|
| Implemented smarter logic for when to switch from distributing rules to distributing words. Implemented faster line skipping when distributing words. |
|
More...
|
over 11 years ago
|
| Implemented --node support for wordlist mode without rules (in addition to with rules, which was added in a prior commit) |
|
More...
|
over 11 years ago
|
| In restore_line_number(), use fgetl(), not fgets(). This makes a difference if there are over-long lines. |
|
More...
|
over 11 years ago
|
| Added logging of the "Node numbers ..." message |
|
More...
|
over 11 years ago
|
| Moved printing of the "Node numbers ..." message from opt_init() to john_load() |
|
More...
|
over 11 years ago
|
