Proposed patch for loop.c

[Toon Moene]

Bernd,

>I've been trying to find out why the strength-reduce pass occasionally
>generates awful code on the i386. First, an example to demonstrate what
>I'm talking about (to the left the "normal" code generated by >egcs-980525, to the right the code that is generated with the patch 
>below).

[ Sorry to send this from my work address, but I'll be away for a couple
  of days - please reply to toon@moene.indiv.nluug.nl ]

This will be addressed in full by the giv-combine patch Richard
Henderson is working on (see the "loop optimizations, mark 10" thread on
the egcs mailing list).

HTH,

--
Toon Moene, KNMI, PO Box 201, 3730 AE De Bilt, The Netherlands.
Tel. +31302206443, Fax +31302210407,  e-mail moene@knmi.nl
URL: http://www.knmi.nl/hirlam