Re: Converting xconq to C++

[Erik Jessen <erik dot jessen at home dot com>]

I've found that simply getting C code to compile in a C++ compiler flushes out
various little bugs, etc.  How about this: let's get the code to compile at all
in C++
compiler first?

I support picking C++ as well, when there are a lot of people mucking with the
code,
objects make it a lot easier to isolate changes.  As long as the person doing
the
original conversion understands how to use C++ ('has a' vs. 'is a' etc.).

I would *really* love a network play version.

Regards,
Erik

Hans Ronne wrote:

> >I'm somewhat cool on C++, with the #1 reason being that g++/stdc++
> >isn't as stable as gcc-for-C yet.  But that's not why I'm writing....
> >
> >> Too bad, really, I've had a network version of XconqPPC lying around
> >> since January that is ready for prime time.
> >
> >Drool, drool, drool.  Did you make some bugfixes which aren't checked
> >in yet?  Or do you think there is something with networked games which
> >is broken for tcltk but not for Mac?
>
> No, the Mac version has the same synch problems. This has to be fixed in
> the kernel. There are some subtle differences between the two which reflect
> differences in how the main event loop works, but most games will fall out
> of synch eventually. I have been able to play simple ones like flattops to
> the end, but that  is exceptional. Some actions, such as turning the AI on
> or off, have to be avoided since they always trigger a synch problem.
>
> What I meant was that everything else (chat window etc) now works fine on
> the Mac. It's all in the CVS sources, but compiling them on the Mac is not
> trivial any longer. You need to get the GUSI library and follow the
> instructions that come with it carefully. I put some further instructions
> at the end of mac.c.
>
> Hans
>
> Hans Ronne
>
> hronne@pp.sbbs.se