Re^3: pure Perl SHA-256 crypt()

Replies are listed 'Best First'.
Re^4: pure Perl SHA-256 crypt() by moritz (Cardinal) on Sep 05, 2011 at 14:30 UTC
crypt() in SHA-256 and SHA-256 as provided by the sha256sum command line program aren't exactly the same. How do they differ? Perl 6 - second systems done right	[reply]
Re^5: pure Perl SHA-256 crypt() by zentara (Cardinal) on Sep 05, 2011 at 14:45 UTC
My understanding of the difference is a digest sum is like a fingerprint when taken on a file, an advanced checksum on the file. Whereas the crypt uses the same digest sum algorithm, but adds a salt value, to make a 1 way fingerprint of a password. I'm not really a human, but I play one on earth. Old Perl Programmer Haiku ................... flash japh	[reply]
Re^5: pure Perl SHA-256 crypt() by Perlbotics (Archbishop) on Sep 05, 2011 at 15:30 UTC
One could say: There is a family of `crypt()` functions parameterised by the embedded function (DES, Blowfish, MD5, SHA-256, etc.) that realises the desired one-way (hash) property. The early version of `crypt()` used DES to simulate a one-way (trapdoor-)function. Since the same password produces always the same `crypt()`-output, it would be too easy to pre-compute a lookup table (rainbow table) for a given `crypt()`-value to reveal the original password. Thus, as zentara said, a salt was added to the input to make lookups harder since that would require bigger tables. AFAIR, there were also some DES-related export restriction issues, that would make a non-DES version attractive. SHA-256 is a digest-function of its own. The SHA-256-`crypt()` uses this hash-function as a replacement for DES (or MD5) - and adds some more rounds, etc. to take into account, that modern computers are faster than 30 years ago. Recently, `bcrypt()` has been introduced using another embedded cryptographic function (Blowfish) with a salt and several iterations. In order to allow various `crypt()`-functions the `$type$salt$crypt` notation was introduced (see `/etc/shadow`). My short CPAN-research yield no implementation of SHA-256-Crypt(), but it should be feasible to implement a (not too fast) pure Perl implementation based on the modules that AM has suggested. Alternatively, it might be more useful to create an XS-wrapper around the reference implementation. P.S. I used the terms hash-function and one-way/trapdoor-function a little bit sloppy, real cryptographers may pardon me that. Update: First pure Perl draft implementation ready. Needs 15s to complete ;-) (slow as predicted)	[reply] [d/l] [select]