Forall bugs

(Posted to list in case other poeple are working in a similar area)
I've just fixed a bug in the forall code where we were incorrectly 
multipying the index into temporary arrays by the element size. This could 
be the cause of  some of the bad code/segfault errors.

It doesn't seem to fix the forall_4.f90 failure at -Os though. We're 
definitely still doing something wrong. Looking at the tree dumps we 
sometimes use temporary variables before we assign a value to them. I 
suspect these are generated by the scalarizer, then not combined in the 
correct order.

We also create & initialize temporaries which are never used although this 
may just be the scalarized being dumb and inefficient, rather than actual 
bugs.

Paul



-------------------------------------------------------
This SF.Net email sponsored by: Parasoft
Error proof Web apps, automate testing & more.
Download & eval WebKing and get a free book.
www.parasoft.com/bulletproofapps1