On Mon, Aug 24, 1998 at 05:38:27PM -0400, Todd Vierling wrote:
> 
> : (insn 577 576 579 (set (reg:SI 149)
> :         (plus:SI (reg/v:SI 86)
> :             (const_int 276032))) -1 (nil)
> :     (nil))
> 
> Hm.
> 
> Does toplev.so compile without optimization? (i.e. is this an optimizer
> bug?)

Actually toplev.c is part of egcs. The source file in question is 
usr.bin/compress/zopen.c, and the line number in question is the final
line of the (quite large) function zwrite.

The same problem occurs with optimization levels 0,1,2,3,4 with the
difference being that the numbers in the parentheses vary with optimization
levels.

The diagnostic is issued by fatal_insn called by fatal_insn_not_found,
which make me suspect that gcc has got a wrong or even garbled description 
of an instruction to generate.


- Erik