On Fri, 2005-02-11 at 11:17 -0800, Dale Johannesen wrote: > This bug (which I just filed) is an ICE in the -ftree-loop-linear code. > It's expecting the > ssa-trees for different index variables in nested loops to refer to > different user variables.
It shouldn't be expecting this. and that's not where the ICE actually comes from. The real problem is that their is a linear offset generated in the transformed loop, and since this never triggered so far (we never made transforms that complex), i never added code to convert complex linear offsets back into gcc expressions. You just need to make lle_to_gcc_expression handle linear offsets, and everything will turn out okay.
