[prev in list] [next in list] [prev in thread] [next in thread]
List: kde-core-devel
Subject: Re: 2.2 RELEASE SCHEDULE (proposal) [NOTE: 2.2 != 2.1.1]
From: Thiago Macieira <thiagom () mail ! com>
Date: 2001-03-05 23:52:03
[Download RAW message or body]
David Faure wrote:
> > So, my suggestion: the code is integrated into the sources, but it \
> > won't be enabled by default for 2.2. Or 2.2 beta 1 at least. It would \
> > be #ifdef'ed out.
>
>
> If you ifdef it out (and require the 'user' to enable the code), then \
> it's not really useful - won't be more tested than right now with a \
> separate patch...
Well, then we leave it enabled. If we find it to be too buggy for the \
release of beta1, we can make it disabled by default till we fix things.
> Doesn't it only matter for IPv6 "users" anyway ? How does that work ?
> Configure check ? Or rather runtime-detection of the type of address used \
> ? The primary goal should be that everything works the same for IPv4 \
> users ;-)
Actually, no. The idea is to replace all gethostbyname() and
getaddrbyname() calls with getaddrinfo()/getnameinfo().
Right now, any systems with getaddrinfo() will work. But if it is
missing or PF_INET6 is undefined, then the code will not compile. I'll
sort this out soon.
The idea behind getaddrinfo() (which is defined in the latest POSIX
standard, as far as I know, and defined in RFC) is that it is
transparent to the socket type being used. So the code doesn't have to
know whether the sockets are IPv6, IPv4, Unix or any other kind.
So, no runtime detection is done by our code. If any is done, it's in
libc or libbind.
Note: my code also adds extended FTP code (EPRT and EPSV).
--
Thiago Macieira - UFOT Registry number: 1001
thiagom@mail.com talha@geocities.com
ICQ UIN: 1967141 PGP: 0x8F2978D5 and 0xEA9037A5
Registered Linux user #65028
96/387. All hands! Brace for impact! -- Cmdr. William T. Riker, Star
Trek: Generations
[prev in list] [next in list] [prev in thread] [next in thread]