Re: Broken su in current - trying to fix myself, help needed!

David O'Brien wrote:
> On Thu, Oct 18, 2007 at 05:36:15AM +1000, Peter Jeremy wrote:
>> This bug (or another one with identical symptoms) affects the 'su'
>> commands within the ports build system if you are using tcsh or zsh
> 
> or bash
> 
> It has been suggested that reverting rev 1.79 of su.c may address the issue.
> 
> ----------------------------
> revision 1.79
> date: 2006/01/03 09:17:04;  author: brian;  state: Exp;  lines: +14 -1
> Fix the other su bug reintroduced two commits ago, namely
> 
>     $ su
>     % kill -STOP $$
> 
> where su is executing (t)csh.  csh's job handling is a little more
> special than that of (a)sh, bash and even zsh and blows up a little
> more spectacularly.  This modification restores the original mucking
> about with the tty pgrp, but is careful to only do it when su (or
> su's child) is the foreground process
> ----------------------------

Well.... if 'blows up...' wasn't hyperbole, back in 2006, and IF calling for a 
pid on a dead/nonexistent process reliably returns the *same* invalid pid tday 
(100000 was cited) it might be safer for the time being to do a compare on that 
value then continue the march if matched.

Belt and braces - at least 'til it can be better checked and cleaned up properly.

JFWIW, pkgdb here lists around 600 packages and on each of three test slices, 2 
i386, 1 amd64.

Probably 90% were invoked from a csh, the rest with bash.

Maybe 80% were built from a terminal in xfce4, and guessing that 80% were 
pkg_add'ed, rest were:

cd /usr/ports/<whichever>/<whatever> make install [-DFORCE_PKG_REGISTER].

Quite a few more attempted and failed did so for clear reasons not related to su.

Not ONE instance of any *detectable* difficulty with su as-it-is-now.

But 100% were built su'ed to root *before* running.

Bill