Re: Buildreport for GnuTLS 2.8.3

[Eric Blake]

> > If others agree that adding this crutch is worthwhile
> 
> Yes, I'm in favour of this new module.

Just because a module is present doesn't mean I have to
use it; so I am not opposed to adding the module now that
others have expressed interest in in (the note in
posix-functions/siginterrupt.texi about recommending
sigaction over siginterrupt is enough of a reminder for
me, anyways).

> 
> > +  if (sigaction (sig, NULL, &act) < 0)
> > +    return -1;
> > 
> > POSIX ignores sigaction failure here, rather than returning -1.
> 
> That's an instance of bad coding practice.

In other words, is it worth trying to write the siginterrupt.m4
macro to detect if any vendors have a buggy siginterrupt
implementation, where copying the POSIX example mishandles
sigaction failure, in order to pick up our better coding practice?
Or is it something that can only be detected with valgrind?

Also, should we try to give feedback to the POSIX folks
suggesting how they can beef up their samples?

-- 
Eric Blake

-- 
View this message in context: 
http://www.nabble.com/Re%3A-Buildreport-for-GnuTLS-2.8.3-tp25008027p25015880.html
Sent from the Gnulib mailing list archive at Nabble.com.