[jira] [Issue Comment Edited] (CODEC-133) Please add a function for the MD5/SHA1/SHA-512 based Unix crypt(3) hash variants

["Christian Hammers (Issue Comment Edited) (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CODEC-133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13243758#comment-13243758 ] 

Christian Hammers edited comment on CODEC-133 at 4/1/12 4:02 PM:
-----------------------------------------------------------------

A new approach: While playing around with the new Kotlin JVM language, I tried to convert the original C sources of MD5 and SHA2 crypt() to Kotlin and after this to Java just to see the differences. The nice benefit of this excersise is that we now have Java implementations that are not only better commented than the ones from UTexas but also sufficiently different to not have any copyright problems. Any resemblance is due to the fact that we both translated the same C code nearly line by line.

So please accept the attached patch "commons-codec-crypt3.diff"! :-)
                
      was (Author: lathspell):
    A new approach: While playing around with the new Kotlin JVM language, I tried to convert the original C sources of MD5 and SHA2 crypt() to Kotlin and after this to Java just to see the differences. The nice benefit of this excersise is that we now have Java implementations that are not only better commented than the ones from UTexas but also sufficiently different to not have any copyright problems. Any resemblance is due to the fact that we both translated the same C code nearly line by line.

So please accept the attached patch :-)
                  
> Please add a function for the MD5/SHA1/SHA-512 based Unix crypt(3) hash variants
> --------------------------------------------------------------------------------
>
>                 Key: CODEC-133
>                 URL: https://issues.apache.org/jira/browse/CODEC-133
>             Project: Commons Codec
>          Issue Type: New Feature
>    Affects Versions: 1.6
>            Reporter: Christian Hammers
>              Labels: MD5, SHA-512, crypt(3), crypto, hash
>         Attachments: commons-codec-crypt3.diff, crypt3-with-utexas-licence.diff
>
>
> The Linux libc6 crypt(3) function, which is used to generate e.g. the password hashes in /etc/shadow, is available in nearly all other programming languages (Perl, PHP, Python, C, C++, ...) and databases like MySQL and offers MD5/SHA1/SHA-512 based algorithms that were improved by adding a salt and several iterations to make rainbow table attacks harder. Thus they are widely used to store user passwords.
> Java, though, has due it's platform independence, no direct access to the libc functions and still lacks an proper port of the crypt(3) function.
> I already filed a wishlist bug (CODEC-104) for the traditional 56-bit DES based crypt(3) method but would also like to see the much stronger algorithms.
> There are other bug reports like DIRSTUDIO-738 that demand those crypt variants for some specific applications so there it would benefit other Apache projects as well.
> Java ports of most of the specific crypt variants are already existing, but they would have to be cleaned up, properly tested and license checked:
> ftp://ftp.arlut.utexas.edu/pub/java_hashes/ 
> I would be willing to help here by cleaning the source code and writing unit tests etc. but I'd like to generally know if you are interested and if there's someone who can do a code review (it's security relevant after all and I'm no crypto guy)
> bye,
> -christian-

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira