[PATCH] Fix a FSM threading ICE (PR tree-optimization/65735)

[Jakub Jelinek]

On Sat, Apr 11, 2015 at 07:19:00PM +0200, Marc Glisse wrote:
> On Sat, 11 Apr 2015, Jakub Jelinek wrote:
> 
> >@@ -1391,13 +1391,13 @@ thread_through_normal_block (edge e,
> >      vec<basic_block, va_gc> *bb_path;
> >      vec_alloc (bb_path, n_basic_blocks_for_fn (cfun));
> >      vec_safe_push (bb_path, e->dest);
> >-      hash_set<gimple> *visited_phis = new hash_set<gimple>;
> >+      hash_set<basic_block> *visited_bbs = new hash_set<basic_block>;
> >
> >      max_threaded_paths = PARAM_VALUE (PARAM_MAX_FSM_THREAD_PATHS);
> >-      fsm_find_control_statement_thread_paths (cond, visited_phis, bb_path,
> >+      fsm_find_control_statement_thread_paths (cond, visited_bbs, bb_path,
> >					       false);
> >
> >-      delete visited_phis;
> >+      delete visited_bbs;
> >      vec_free (bb_path);
> >    }
> >  return 0;
> 
> I understand minimizing the patches right before the release. At any other
> time, it would have been a great occasion to remove this new/delete
> anti-pattern.

Not at this point, I really wanted to do RC1 on Friday, now it looks more
likely for Monday, but really only blocker bugs at this point should go in.

	Jakub