Re: Was: Rare Bug Now: Dumping Cores

From: Mike Stilson (mike@c746148-a.ehlls1.pa.home.com)
Date: 10/08/01


On Tue, Oct 09, 2001 at 11:11:04AM +1000, Caniffe wrote:
>Hello,
>
>> Or set it w/in the program itself.
>
>Granted, could be done - but I question why you'd want to wrap it in
>Circle's code itself.  IMHO, if you're developing a mud, chances are
>you're probably developed or are developing something else at the same
>time (or in future), and as such, are you really going to want to set the
>limits internally in code for each program you may write?
>
>Take CircleMUD's utilities.  What if they need to dump core?  Having
>limits set in Circle's initialisation won't help these one iota.

I don't bother doing it in everything, of course not.  limits are set
variously by pam or login.  However, the mud runs as its own user/group
with its own constraints.  Occasionally the problem crops up of starting
it manually (should autorun crap out for some reason, the moon aligned
wrong with pluto, who knows), and it would pick up the limits of
whichever person restarted it, occasionally low enough to prevent it
dumping core (Sometimes I even turn mine off when testing certiain
programs that I don't want 10M+ core files lying around that I won't
look at).

That all said, the ulimit -c unlimited method is far and away the
easiest, though.

-me

--
   +---------------------------------------------------------------+
   | FAQ: http://qsilver.queensu.ca/~fletchra/Circle/list-faq.html |
   | Archives: http://post.queensu.ca/listserv/wwwarch/circle.html |
   +---------------------------------------------------------------+



This archive was generated by hypermail 2b30 : 12/06/01 PST