[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Tricks for Windows coding and win version status

From: Ingo Ruhnke
Subject: Re: Tricks for Windows coding and win version status
Date: 08 Aug 2002 01:04:02 +0200
User-agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.2

"Alberto Curro" <address@hidden> writes:

>   First of all, only say that I've no changes from my last mail,
> time ago. The Windows version still goes on compiling, but I cannot
> make it running... and I've already tested 10 different versions
> from Win98/98SE/Me to WinXP Home/Pro, different DirectX versions,
> hardware, PCs, etc. The main error (click on Play and the game stops
> running) is still here.

I havn't yet discovered a single bug in Pingus that was caused by the
OS, DirectX version or other rather abstract things. There are
basically only two kind of errors, errors in my code or errors in
ClanLibs code, so focusing on these kind of errors is important to not
get into a try&error situation and search in the wrong location.

> I'm thinking about asking for help to Clanlib people, as I cannot
> doing anything simmilar to a "real" debug with the debug libraries
> from Clanlib (both sources and binaries tested) and I have no
> startpoint to work on.

Not sure on how to get the debugging version to work, but in ~80% of
the cases a bunch of std::cout is enough for the debugging, especially
if you do not get a crash, but just something not working as expected.

> The latest changes on the sources seems to have solved some of the
> other problems and now the game seems to run better (or, at least,
> it starts faster and I can let it running in the main menu with no
> crashes, that was impossible before), but not this one.

The whole input and gui system was basically rewritten from ground up
and the rewrite is still going on in a few places.
>   Ingo, any idea about any change from 0.4.x to this release that
> should have this effect? I can still run that version, is slow and I
> have no sound, but it works.

0.4.0 is very very old (~2 years?!) and basically there are hundreds or 
of things that changed since then.

> Any other idea to test?

Does it currently crash or just exit at an unexpected point? If the
later is the case, inserting a ton of std::cout here and there in the
code might make it easier to track down where it is falling and
causing the incorrect behaviour. For crashes, you might need to
compile the debug versions, since std::cout's won't help very much at

>   First, the most common point : due to some lack of bug in the
> mannage of namespaces from VC++, I have a real headache with
> hundreds of warnings since VC++ seems to mix namespace's names, and
> it warms about a namespace with a name length bigger than 255
> characters.

There is IIRC some #pragma to switch these warrnings of, ClanLibs code
has a few of them (maybe this one #pragma warning( disable : 4786)?!).

>   Visual C++ seems to get confused when you have declared new
> namespaces and use also the STL ones. This is clearer when you do
> something like this : [...]

Not sure about that one, #pragma might help. Could you post the
warnings that you get?

Advent:         || Ingo Ruhnke <address@hidden>
Feuerkraft:     || 
Pingus:           || ICQ:      59461927
Home:  || JabberID: address@hidden

reply via email to

[Prev in Thread] Current Thread [Next in Thread]