Endianess

[Toon Moene]

>  No!  This is a FAQ, this is true, but it isn't a egcs
>  FAQ.  It is a how do I program in C FAQ.  We don't
>  maintain that FAQ.  The reason why we cannot list (chouls
>  not list) this in our FAQ, is then we must have all
>  answers to all programming FAQs, including how to program
>  in C, C++, Fortran, Ada, Objective-C, you get the point.
>  The we must replicate all those preexisting FAQs into
>  our, and replicate the work, and do a poor job of it,
>  because we are as interested in maintaining it and so
>  on...

Mike, I agree completely with you - the reason I tend to answer  
these questions - irrespective if they're issued to fortran@gnu.org,  
egcs-bugs@cygnus.com or comp.lang.fortran is that it probably took  
the original questioner a lot of time to find out what was wrong.

By the time a Fortran programmer like Mr. zhumin finds out that the  
actual problem he's experiencing is caused by an endianness  
problem, he already has expended several hours of pointless  
debugging and ploughing through manuals.  I speak from experience,  
and it's only because I have a hard-coded table in my head that sums  
up the wrong-endian architectures (like Intel and Alpha) that I can  
proceed with some speed in this business.

Yes, I wish I could keep this issues out of our mailing lists -  
unfortunately, I don't know of a 100 % sure method.

Cheers,
Toon.