[JDEV] Re: jabberd 1.4.3 release candidate again

[Frank Seesink]

Matthias Wimmer wrote:

...
> AFAIK inet_ntop() is not implemented in bind but in the glibc. The
> interesting part of this is that inet_ntop() is part of native Windows
> as far as I know. I'm surprized that it isn't supported by Cygwin.
> 
> I don't know how it should be replaced. Sure you can implement the
> function yourself, but the older function inet_ntoa() is not really a
> choice as it is very limited in functionality and I'd call it depricated
> ;)

Hey Matthias!

Thanks for the feedback.  Since I haven't done C coding in a good bit, 
and I've never been a "professional" C coder (as in "sit at a desk and 
code for my paycheck"), did a little Googling.  Though I agree that it's 
strange, from what I can tell, Cygwin does not support IPv6 at this 
point.  The best quote I found, from one Corinna Vinschen, Cygwin 
Developer, Red Hat Inc., dated 2 Apr 2002 on the Cygwin mailing list:

	"No INET6 support in Cygwin so far."

But haven't found anything more recent than this to indicate a possible 
change.

I'm running the latest Cygwin (1.5.5-1) with _all_ the latest packages. 
  I did this on purpose so I would every possible tool at my disposal 
while figuring out how to get Jabberd working, and I had the HD space to 
burn so what the heck. :-)  Looking in /usr/include/arpa/inet.h, it 
seems only the older inet_ntoa() is an option.  There are no 
declarations indicating IPv6 support.

So I guess that kinda kills that, at least in terms of quick/simple 
fixes.  Hey, I realize I'm fighting an uphill battle, and it's more 
hobby/tinkering than anything (it's been great for refreshing my coding 
skills :-) ).  I'm taking a *nix-based project and hacking it under a 
pseudo *nix environment which sits on top of a very NON-*nix 
environment.  Yeah, they have some Unix concepts in NT/2K/XP, but all 
the worst parts if you ask me.

And the mindset involved in programming often is heavily influenced by 
the platform under which development is done.  Unix programming is often 
based on taking lots of little pieces and building them into a final 
product, often using bits & pieces already in the OS to save you 
time/energy (why wouldn't you?)...and counting on those bits & pieces to 
already BE there.  Windows development, though you have the APIs, tends 
to follow a rather different approach.

Let's face it.  Projects like this are really designed to build/run 
cleanly under *nix.  And if I were setting up a serious IM server, I 
would most definitely NOT use Jabberd or PostgreSQL under Cygwin.  I 
would build myself a proper Linux/*nix box.  But these projects are fine 
for small environments where dependence on them is limited...if we can 
get them to work. :-)  And besides, in a masochistic sort of way, it's 
been fun getting the pieces I have to work.  [I have yet to tackle the 
transports though...ugh.]

</rambling>