On Fri, May 19, 2006 at 02:01:02PM +0100, Alex Bligh wrote:
> I have some code which at least on my machine allows the user to
> continue after a SEGV. I hope this will be in a check-in-able
> state this evening.
>
> I wouldn't normally check this in so close to 0.5, but then it
> only ever gets called if there is a SEGV, and as that currently
> means "dialog box and goodbye", it can hardly make things worse.
It can.
>
> If I am reasonably happy it is stable, do you want it?
>
If the SEGV signal did not originate from a kill(2) invocation the
program behaviour after catching it is undefined.
You certainly do not want this. It can lead to infinite loops, memory
exhaustion or even data damage and loss.
SIGSEGV cannot be over-lived: the app should better quit just
like any other app does.
--
Vasil Dimov
gro.DSBeerF@dv
Testing can show the presence of bugs, but not their absence.
-- Edsger W. Dijkstra
Attachment:
pgpiztwCJ2rq5.pgp
Description: PGP signature