Re: [patch 1/4] Separate gimple.[ch] and gimplify.[ch]

2013-11-12 Thread Andrew MacLeod
On 11/12/2013 02:23 PM, Jeff Law wrote: On 11/11/13 14:03, Andrew MacLeod wrote: So on to the anomaly that causes the issue. force_gimple_operand* is extensively used by the middle end, but not by the front end at all. The front ends do not use the statement iterators, but 2 of the force_gimple

Re: [patch 1/4] Separate gimple.[ch] and gimplify.[ch]

2013-11-12 Thread Jeff Law
On 11/11/13 14:03, Andrew MacLeod wrote: So on to the anomaly that causes the issue. force_gimple_operand* is extensively used by the middle end, but not by the front end at all. The front ends do not use the statement iterators, but 2 of the force_gimple_operand() routines work with gsi's, and

[patch 1/4] Separate gimple.[ch] and gimplify.[ch]

2013-11-11 Thread Andrew MacLeod
This was/is relatively painful, and there will be another set after this to really finish it. The front ends and the middle end both make extensive use of gimplification There is no way at this early point to really separate them unless we created a gimplify-fe.[ch] and gimplfy-be.[ch] so a