Iterated-sha1: Don't accept raw ciphertexts except XSHA

[magnumripper]

Bare XSHA is exactly 48 chars of uppercase hex. Pot entries will currently be iterated-sha1 canonical, tagged and in lowercase.

I guess we should write XSHA entries to .pot using the raw uppercase format, like the XSHA-specific CPU format does? This will complicate valid a little and need adding a split() but it's trivial. Working on it (slowly).

[magnumripper]

I guess we should write XSHA entries to .pot using the raw uppercase format, like the XSHA-specific CPU format does? This will complicate valid a little and need adding a split() but it's trivial. Working on it (slowly).

Done. Actually no split() needed as we require uppercase for XSHA and lowercase for others. I could even drop prepare().

[solardiz]

I agree with the changes here.

However, I think our checks for the format tag should be case-sensitive. I now see we also have similar case-insensitive matching of tags in a few other formats - those are probably also wrong in doing that - but let's at least not make matters worse by replicating this in new formats without a reason.

[magnumripper]

However, I think our checks for the format tag should be case-sensitive. I now see we also have similar case-insensitive matching of tags in a few other formats

Wow, I never noticed. I'll fix this one but we should probably fix all other formats (giving each some thought - there might be some where it's sensible although I can't think of any).