Re: [Pipmak-Users] Problems with compiling Pipmak with Microsoft Visual C++

[Christian W. <cwa...@gm...>]

Andrea Viarengo wrote:
> So I have tryed to compile Pipmak from SVN with MS Visual C++ 2005 Express.

Nice! That's very welcome. I've only compiled Pipmak with GCC so far, 
and getting it to work with other compilers could be beneficial for the 
code quality. (Even though I've been using VC++ 6 at work for almost a 
year now, I've never tried to compile Pipmak with it.)

> If I run pipmak without a project, I get only a window named "pipmak" 
> without the contents (I can see the windows/desktop below),
> and pipmak loop forever on function "enternode()".

I'm not sure what you mean by that. There's no function "enternode()" in 
the C code, do you mean it calls the Lua function you specified using 
"onenternode" again and again? Or it actually does an infinite loop in 
Lua code where it shouldn't?

> Do you think the problem was on PhysicFS? 

Looks like it has trouble reading ZIP files. You should be able to work 
around this by unzipping the "Pipmak Resources" file and placing the 
contents in a folder named "Pipmak Resources" (or making a copy of the 
pipmak/resources folder from SVN named "Pipmak Resources").

> I think that PhysicFs was used also to get resource from Pipmak resource,
> isn't it?

Exactly. So, you also don't see the hand cursor? (What else in place of 
it?) Actually, it's a surprise that anything works at all if the Pipmak 
Resources archive can't be read, because a lot of functionality is 
implemented in resources/defaults.lua in there.

> Where is defined M_PI, I didn't found it (For now I add M_PI=3.1415 in
> the compiler command line)

Should be in <math.h>. Did I forget to explicitly include that somewhere?

> There are a lot of warning about casting types in a small types, 
> like:
>    double to float
>    int to GLfloat
>    double to GLfloat

These should be harmless in practice, but if you can tell me where they 
are I can look into fixing them. Most are probably unintentional, and 
there is a performance penalty in unneccessary conversions.

> and incompatible type:  char[xxx] and 'LPCWSTR'

That looks more serious. 'W' in LPCWSTR sounds like "wide", i.e. 
Unicode, and that's probably not what we want. Where does this occur? I 
don't think I use such Windows-isms as LPCWSTR anywhere in my code...

  -Christian