[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: pam_{unix,pwdb}: crypt/md5 necessary?



Andrew Morgan wrote:
> 
> That may well work.
> 
> One caveat, is that when first implemented we goofed on bigendian
> support and for legacy reasons now support both endian forms of the md5
> algorithm (one is not really md5 as a result). I doubt glibc will
> correctly handle those passwords.

Oh, this explains that strange two "good" and "broken" crypts in those
modules.  Glibc will _not_ handle this.  But may I ask -- when that mistake
was made and how many time was before it was corrected?  Maybe it can
be gone now?

> If you want to try this and report your findings/submit a patch:

Ok, I'll do this.  But some issues before (already sent in a separate mail)...

>  http://sourceforge.net/patch/?group_id=6663

Does Sourceforge requires to log in/register to post a patch?  I don't like
that very much -- any time I registered somewhere I started to receive SPAMs,
especially if I do something "informative" there...  But it is not an issue..


Regards,
  Michael.



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index] []