arm.[ch] tidy ups.

[Richard Earnshaw]

nickc@cygnus.com said:
>   I am going to check in the patch below which fixes a few compile
>   time warnings and errors when building gcc for arm-pe.
>
>   One possibly interesting thing is that I have created two new macros
>   HOST_INT and HOST_UINT which can be used to specify host sized
>   integers.  This not only saves space in the sources but also allows
>   us to eliminate the compile time warning about the behaviour of
>   unsuffixed integer constants under K&R compilers. 

I'm not convinced this is the right way to fix this problem, unless we are 
going to do it across the whole of the compiler.  Surely this isn't a 
problem that affects only the ARM backend?

I really don't like the fact that gcc -Wtraditional gives a warning for

	unsigned long x = (unsigned long) 0xffffffff;

How else is one supposed to write this in a way that is acceptable to both 
K&R and ANSI compilers?  How else could it be interpreted?

If we are going to leave -Wtraditional on, I think this is a general 
problem that should be fixed in a common way across the whole compiler 
source tree.

R.