Re: renaming core dumps

Linus Torvalds (Linus.Torvalds@cs.helsinki.fi)
Sat, 24 Feb 1996 11:00:47 +0200


"Carsten Paeth": "Re: renaming core dumps" (Feb 23, 9:46):
>
> It's amusing, one or two (?) years ago, we already have had core.command in
> an alpha kernel. After the discussion Linus take it out again.
> I hope this discussion will have the same results, 'cause I don't want it.
> But I have no problem with a config option.

Actually, the code is _still_ there, but it's behind an "#if 0".

Essentially, I don't think I'll change the core-file naming, for the
following reasons:

- it buys you very little
- it's non-standard and lots of people didn't like it.

Most people have probably noticed that core-files are disabled
completely by default by the kernel (core size limit=0), and that kind
of gives you a hint about my opinion of core-files. It is not that
core-files are useless (far from it), but core files are rather special,
and you don't want them unless you're debugging a program that you
already know is buggy.

(Having core-files enabled all the time on the assumption that
"something could happen" is a broken setup, imho. If your system is
that unstable you have other problems).

And if you do enable core-files, it very seldom makes much sense to have
them named differently. In my opinion a core-file is special enough
anyway, that if you want to save it you might as well re-name it by
hand.

Anyway, I could live with a config option too, but I do not really see
the reason for it. The fact that gdb dumped core on somebody is
unfortunate, but not a catastrophy (you can just re-do the core-file,
after all, total time loss 1 minute).

Linus