Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ames!uhccux!munnari.oz.au!murtoa.cs.mu.oz.au!otc!metro!extro!glenn
From: glenn@extro.ucc.su.oz (G. Geers [ext 3241])
Newsgroups: comp.unix.xenix
Subject: Re: Xenix Compiler
Keywords: Compiler Error
Message-ID: <597@extro.ucc.su.oz>
Date: 10 Sep 89 23:47:33 GMT
Reply-To: glenn@extro.ucc.su.oz (G. Geers)
Organization: University Computing Service, Uni. of Sydney, Australia.
Lines: 43

In article <1024@polari.UUCP> corwin@polari.UUCP writes: "... getting a 
'Compiler Error (Internal) : infinite spill'. ..."

I have had this error message and like you I could not find any documentation.
However, have a look on the line indicated and you will probably find soemthing
like:
		a = foo[a++];

or perhaps the assignment of a struct to an array of structs using an element
of the struct as the array index, e.g.

typedef fred struct {
			int a;
			char *p;
			};

			fred *f;
			char ad[20];

			...

			f->p = ad[f->a++];

			...

This can be rectified by judicious use of temporary variables. Note that this
doesn't happen in small model where sizeof(int *) = sizeof(int). I have *never*
seen it on our 386.
The real problem is when this happens in the control variable of a for loop -
it can be a real pain.

I wonder if you could send me the code fragment or tell me what program you
are playing with (if its pd I might have it and the fix).

The problem also occured much more frequently with 2.1.X DS.

	Hope this helps,
				Glenn

glenn@extro.ucc.su.oz

p.s. Have you read "Nine Princes in Amber" ? If yes then maybe you can tell me
	if the 4th book in the 2nd series has been released yet.