On Wed, 2005-03-23 at 13:32 -0800, Bill Kendrick wrote:
> As mentioned, we have stuff like this, now:
> #if !defined(_SDL_TTF_H) && !defined(_SDLttf_h)
> which seems to cover the schizophrenia that the SDL header files have been
> going through as they mature. ;^)
> Does THAT not work?
No. It's a case of bit-rot. The Tuxpaint in Debian doesn't have that.
Someone tried to build it against a recent SDL, and it broke. I'm sure
the code above works just fine. I just haven't done a CVS build in a
> > > The proper way to do this would be to have a configure script
> > > that could then give an error message.
> Ack! No! ;^)
> Or, if so, whoever creates it will need to become adopted by my wife and
> I and live in our spare bedroom, so that when stuff stops working, I'm
> not stuck. ;^)
Hah! OK. I concede. I wouldn't want it to come to that.
> Anyway, if it comes to it, I suppose we can just remove the checks.
> I was TRYING to be friendly, but SDL*.h went and broke things out from
> underneath me. ;^)
So they did. And that is my final word on it, if anyone asks. "Sorry,
the build system is correct, it's upstream that broke." :) (Too bad the
bug is closed now, or I'd add that!)