cculianu / brute38 Goto Github PK

The DecryptWithPassphrase method takes the key as a string and decodes it to a byte array. This decoding is done with every passphrase guess.

I think performing the decoding once and handing over the byte array as the key will increase performance.

Tested on arm, x86 Linux, Windows, Mac... when running a command with custom characters over a certain amount of custom fields - the additional range of characters does NOT increment. I've tested both with a --resume flag and with just an example. For instance:

~/go/bin/brute38 --charset='ABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890' '?????????????????' 6Psdgsdgsgsdg.... will output to:

KeyType: ECMultKey Network: Bitcoin Pattern: ????????????????? Unknown chars: 17 Password length: 17 Total passphrase space size: 18446744073709551615 1998934881474191231 passphrases tried (latest guess: KJHT3K0JL1GPAAAAA )
^---------------from four characters at the end.
Also the counter, does not appear to count base 36 correctly as you can see from the base 36 character set.

Not sure what to to fix it, I have a wrapper, but this is an interesting limitation I wish I had realized was an issue a few months ago.

Would be nice to support masks for the given character set - to specify a special character, an upper, lower, number etc.

This would be quite powerful when combined with using the pack statsgen.py too to create a mask that could cover a given range.

the default number of threads should be set to the number of physical cores, not logical cores. If you have a 4 core machine with hyperthreads, the default number of threads is 8. This is actually a bit slower than if 4 threads were used, since each core is already fully utilized and the extra 4 threads incur extra context switching.