On 25 Mar 2006 00:02:43 +0000, Gaius Mulley <[EMAIL PROTECTED]> wrote: > Pragmatically I guess it is best for me to maintain a reversed patch > which can be applied to a gcc-4.1.0 tar ball which reintroduces this > TYPE. Any thoughts?
I think it would be better if you make the SET_TYPE a front-end specific tree node, much like e.g. the tcc_type tree codes in cp/cp-tree.def. Then you can use SET_TYPE in the front end, and translate the set operations to valid GIMPLE later on when gimplifying. Gr. Steven
