[3423] in testers
Re: Bugs in attach -Padd
daemon@ATHENA.MIT.EDU (Craig Fields)
Sun Apr 12 21:02:04 1998
Date: Sun, 12 Apr 1998 21:01:58 -0400
From: Craig Fields <cfields@MIT.EDU>
To: ghudson@MIT.EDU
Cc: testers@MIT.EDU
> That may have been intentional, but I think some Bourne shell users
> would like to be able to use athena_path. (I don't know *why*, but
> at least Aaron seems to want to.) At any rate, the documentation
> doesn't mention this gap at all.
Ya, that would be a doc bug. I was unwilling to contemplate what it
might mean for someone to want that. I guess it's no big deal to add
it now that I've been forced to think about it though. :)
> I got "attach -Padd" to seg fault, by running
>
> attach -Padd -e -b netbsd
>
> as root on a Sun.
Sigh. This is partly a doc bug, since that's totally illegal syntax
and the man page doesn't say anything about it. If -e is specified,
then the current values of athena_path and athena_manpath must also be
passed on the command line. Since they weren't (and this condition
isn't also verified), NULLs got passed all over the place. I don't
know quite how I overlooked documenting the -e option for the new
changes.
Anyway, I'll send in patches for this stuff in the next couple of
days.
Craig
