Feature req [masks]: add increment backwards flag

[ysrn]

Currently when using the increment feature it goes from the start of the pattern to its end. I think it would be super useful to have a flag to reverse that direction. I'm hitting up against this use-case a lot, and thought it might be useful to others as well.

Example:
If I'd like to run through all masks starting with a single digit and ending with 2-4 lowercase, that's easy: -i ?d?l?l?l?l

Now, if I'd like to run through all masks ending with a single digit, but starting with 2-4 lowercase. I'm forced to verbosely specify each mask in order: -i ?l?l?d then -i ?l?l?l?d then -i ?l?l?l?l?d.

If I had --increment-backwards, then I could use a simple -i --increment-backwards ?l?l?l?l?d.

[Chick3nman]

While this isn't a rebuttal you could work around this currently by specifying the masks in a mask file to avoid having to run multiple attacks. An "inverse increment" or similar flag could still prove useful.

[dmgarry]

Just thinking about this... May or May not be possible as I can't try it at the minute but if you stdout... put your mask in reverse so last character first and so on... increment it the way you want... pipe it back into hashcat applying the reverse rule?

[Chick3nman]

You could do that @dmgarry but you lose the speed/efficiency of running the mask attack from within the kernel. By using mask files you achieve the same thing but without the potential speed loss.

[PenguinKeeper7]

Please test #3960 and see if it meets what you want. Here's an example command, using your listed situation of an unknown amount of letters, then one number at the end

> ./hashcat -m 99999 abcd1 -a 3 ?l?l?l?l?l?l?l?d -i --increment-inverse --quiet

abcd1:abcd1

[ysrn]

Woah that was fast! I tested it and played around a bit. It does exactly what I described/needed and will be incredibly useful, thank you!

[ysrn]

Found a small bug by looking at the code. Using an escaped question mark ?? in the pattern will lead to it being misinterpreted as a pattern token instead of a literal. So the pattern ?l??1?d will error twice with "Custom charset 1 is undefined" in inverse mode.

[PenguinKeeper7]

Found a small bug by looking at the code. Using an escaped question mark ?? in the pattern will lead to it being misinterpreted as a pattern token instead of a literal. So the pattern ?l??1?d will error twice with "Custom charset 1 is undefined" in inverse mode.

Thanks for the report! I've fixed this issue with cbc3470 and a mask file bug with custom charsets.
There are no known issues so please let me know if you do find another