Front page | perl.perl5.porters |
Postings from September 1999
Re: deprecating SIGDIE
Thread Previous
|
Thread Next
From:
Ilya Zakharevich
Date:
September 30, 1999 11:57
Subject:
Re: deprecating SIGDIE
Message ID:
19990930145705.A27323@monk.mps.ohio-state.edu
On Thu, Sep 30, 1999 at 12:37:28PM -0600, Tom Christiansen wrote:
> >And endless programs/modules which rely on the current
> >well-established semantic of __DIE__. If you want a pseudosignal
> >__FATAL__, call it like this.
>
> The current behaviour is broken and wrong, and Larry has said this.
I repeat: do not put words in Larry's mouth. If he wants to say
something, he can.
I remember that last year he expressed some *opinions* on how things
would better work, but IIRC his concerns were quieted by the
die if $^S; # Or however we write it.
> The question is deprecation, disabling, or substitution.
The question is that Perl's exception handling needs improvement.
However, numerous discussions did show that such an improvement will
not come from one cavalry strike: a well-designed campaign is needed.
On the third hand, I agree that until such a campaign is designed,
some minor fixes to the current scheme may be useful too. I would not
object to having a bastardized __DIE__ which you propose, as far as
the currect behaviour is available too: it is very useful in many
situations, such as debugging/logging. What about __EXCEPTION__ for
the semantic of "old" __DIE__?
Ilya
Thread Previous
|
Thread Next