On Thu, 2014-02-13 at 15:39 -0500, Bruce Momjian wrote:
> On Thu, Oct 10, 2013 at 08:42:14PM -0400, Bruce Momjian wrote:
> > On Thu, Oct 10, 2013 at 08:32:32PM -0400, Bruce Momjian wrote:
> > > > > How can the later entry not be MD5 hash?
> > > >
> > > > Because what you pass to the functions is 'md5', not 'md5 hash', which
> > > > is what the new text appears to indicate.
> > >
> > > So if we revert, will it still be clear what is MD5 and what is MD5 hash?
> >
> > I mean, will it be clear what is MD5 crypt and what is MD5 hash?
>
> I have made the attached doc change, which places "(hash)" outside of
> the literal text block. You are right the literal blocks should match
> what is passed.
I think this entire line of patches should be reverted from the 9.3
branch, because it's not a bug fix, and arguably makes no sense.
I also think that these patches should be reverted from the master
branch, because they make no sense and don't actually address the bug
report. Adding an output length column might make sense as an
independent patch.