[Top][All Lists]

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

Re: [Adonthell-devel] Endianness (Adonthell and Solaris)

From: Jayson Baird
Subject: Re: [Adonthell-devel] Endianness (Adonthell and Solaris)
Date: Mon, 18 Feb 2002 11:14:02 -0600
User-agent: Microsoft-Entourage/

Having worked with code on OS X, Solaris, and x86, the only times you run
into endian-ness are with pre-compiled binaries. Essentially if you say:
N = 4;
N >> 1;
N would then be 2, no matter what architecture. Now on some of the SDL
stuff, you have to always check thinks like
If(SDL_BIGENDIAN) for computing things like masks, but I've never heard of
any problems with simple things like that getting all tchocky.

To the graphics engine people: when creating surfaces (using
SDL_CreateRGBSurface) are you checking for endian-ness? Especially where you
set the ARGB masks?

Just a thought...


On 2/17/02 4:01 AM, "Kai Sterker" <address@hidden> wrote:

> I've been exchanging mail with a guy who tried to run Adonthell on
> Solaris, and there are a number of problems.
> The first one are the precompiled python scripts. For some reason, Python
> does not like them on his box, even though they were compiled on it. (He
> said that he tried different versions of Python, so possibly something
> went wrong there.)
> Anyway, running with the -g option and the data-tarball worked, sort of.
> When it gets to loading the title graphics (the jewelbag, etc ...) there
> are some errors. (He wasn't more specific though.) I imagine that this is
> a matter of byteorder, as I dimly remember that the Sparc is a big-endian
> system, whereas the x86 is little-endian. Have no confirmation yet, but
> it's the only explanation I have.
> Fortunately, SDL provides all the means to work around this. So once I get
> Brian's reply, I'll add the necessary code to convert the data after
> loading, in case the byteorder differs. Shouldn't be a big deal.
> Should that fix the problems, we'll have to make sure to do this
> byteorder-testing (and swapping if neccessary) everywhere where data is
> saved and loaded.
> Btw, the same problem would occur with a port to MacOS X, as the PPC is
> definately big-endian.
> Kai
> _______________________________________________
> Adonthell-devel mailing list
> address@hidden

reply via email to

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