[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [MiNT] Errors while shutting down
- To: mint@fishpool.com (MiNT mailing list)
- Subject: Re: [MiNT] Errors while shutting down
- From: Guido Flohr <gufl0000@stud.uni-sb.de>
- Date: Tue, 12 Oct 1999 22:17:00 +0200
- In-reply-to: <Pine.MNT.4.10.9910121053340.80-100000@rakastaja>; from Martin-Eric Racine on Tue, Oct 12, 1999 at 10:57:21AM +0300
- Mail-followup-to: mint@fishpool.com (MiNT mailing list)
- References: <Pine.MNT.4.05.9910122033470.92-100000@tos> <Pine.MNT.4.10.9910121053340.80-100000@rakastaja>
- Sender: owner-mint@fishpool.com
On Tue, Oct 12, 1999 at 10:57:21AM +0300, Martin-Eric Racine wrote:
> On Tue, 12 Oct 1999, Mario Becroft wrote:
>
> > I noticed a lot of messages scroll past just before the
> > machine rebooted which were too quick to read but I managed
> > to catch some words which read something similar to "OPEN
> > FILEPTR detected."
> >
> > I suppose that this is not the expected behaviour.
>
> This is caused by applications that forget to close all currently
> opened files before exiting, when they receive the KILL signal.
> ...as far as I can tell, anyhow.
Definitely not.
You only have to close files (not FILE* streams, they are buffered) when
you're afraid to run out of descriptors (handles). It is perfectly
alright to exit without closing open descriptors (and btw, freeing memory
on exit is also a waste of time).
In signal handlers it is mostly a bad idea to close files (because that
may block and could trigger other signals/race conditions...). Instead
of closing files you should rather delete them if you haven't finished
writing to them.
On the other hand, if you install a handler for the KILL signal, all of
this is alright, it will do no harm, I can guarantee that. ;-)
Ciao
Guido
--
http://stud.uni-sb.de/~gufl0000/
mailto:gufl0000@stud.uni-sb.de