Re: cvs: php4 /ext/standard file.c
| From: | Sascha Schumann | Date: | Sat, 23 Dec 2000 17:56:56 +0000 |
| Subject: | Re: cvs: php4 /ext/standard file.c | ||
| References: | 1 | Groups: | php.cvs |
| Request: | Send a blank email to php-cvs+get-3281@lists.php.net to get a copy of this message | ||
On Sat, 23 Dec 2000, Zeev Suraski wrote:
> I think that in the last couple of letters we've gone what's reasonable to
> expect in the real world, and it became an academic discussion. We should:
> (a) Start using full_write (we can use my_write() off the shelf, I think).
> (b) Catch this signal, and display a PHP error in case it happens, if we're
> inside PHP's scope.
> (c) Optionally, invoke the original signal handler if we're hit by a signal
> while outside of the PHP scope. If under some circumstances that won't
> work, tough luck. We can give this option up completely too, as I haven't
> yet seen any other modules that pay any attention to this signal.
Have you figured out a way to savely play with other signal
handlers, under the presumption that we cannot know whether
they were set using signal() or sigaction()?
> > Then we are running into other problems, as we don't know
> > whether other modules use signal() or sigaction() to set
> > their handler. A signal handler set using signal() is not
> > necessarily available through the modern sigaction()
> > interface and using signal() to set our own handler won't
> > properly interoperate with handlers which were set using
> > sigaction().
- Sascha