login.c fun

[Rob Landley]

On Monday 11 September 2006 5:29 am, Bernhard Fischer wrote:
> On Fri, Sep 08, 2006 at 07:35:34PM +0200, Denis Vlasenko wrote:
> >Small voyage across login.c code.
> >
> >Code per se is not that bad (I've seen much worse)
> >and applet adds only 4k. However reviewing/fixin it
> >was humerorus enough to share it with the list.
> >
> >(Undoubtedly in the process of fixing it I possibly
> >added my own bugs...)
> 
> The quality of tinylogin (or however it was called) is suboptimal, i
> agree. :)

Yeah, login.c sucks rocks.  Go for it Denis!

Rob

P.S. I have a big stalled passwd rewrite to completely reimplement the shadow 
support in our tree without bolting on a bloated external library of dubious 
licensing, plus using our in-tree support for md5sum and sha1sum rather than 
relying on libcrypt to do that, and making old 3des password support optional 
and thus the dependency on libcrypt optional.

But that's stalled and will have to be adapted to the new reality of the tree 
when I pick it up again anyway (which won't be before bbsh can fully replace 
lash), and I don't think I did any real cleanups to login.c like this.  So 
don't worry about it.  I have work to do in this area _after_ you come to a 
good stopping point. :)
-- 
Never bet against the cheap plastic solution.