develooper Front page | perl.perl5.porters | Postings from April 2006

Re: Making IO::Socket pass test on Win32

Thread Previous | Thread Next
From:
Andy Dougherty
Date:
April 4, 2006 07:31
Subject:
Re: Making IO::Socket pass test on Win32
Message ID:
Pine.SOC.4.62.0604041009100.13709@maxwell.phys.lafayette.edu
yOn Mon, 3 Apr 2006, Nick Ing-Simmons wrote:

> Andy Dougherty <doughera@lafayette.edu> writes:
>
> >When Windows fakes fork(), then I think it *should* have this variable 
> >set.  
> 
> I disagree. Windows doesn't fake fork() in a POSIX sense at all.
> Instead it makes perl's C<fork> operation do almost the right thing.
> But it doesn't create a new process or new memory space or ...

Ahh, I see.  That's indeed different and could well matter in some cases. 
Hmm.  Perhaps the Windows build process should then define something like 
$Config{d_pseudofork} whenever this feature is available.   This would be
documented in Porting/Glossary (which gets carried over to the Config pod 
during the build process).

My main idea was that user code shouldn't normally have to go rummaging 
through $Config{ccflags} to try to figure out what features might or might 
not be enabled.  User code should be able to consult %Config.  I had hoped 
that $Config{d_fork} would suffice, but alas, it won't.

Good catch.  Thanks.

-- 
    Andy Dougherty		doughera@lafayette.edu

Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About