On 10/30/13 03:26, Richard Biener wrote:
I'd assumed that'd be in a follow-up patch and probably would be bounds on the return value or NULL if there were no bounds.diff --git a/gcc/gimple.c b/gcc/gimple.c index 3ddceb9..20f6010 100644 --- a/gcc/gimple.c +++ b/gcc/gimple.c @@ -174,7 +174,7 @@ gimple_build_with_ops_stat (enum gimple_code code, unsigned subcode, gimple gimple_build_return (tree retval) { - gimple s = gimple_build_with_ops (GIMPLE_RETURN, ERROR_MARK, 1); + gimple s = gimple_build_with_ops (GIMPLE_RETURN, ERROR_MARK, 2);Ick - you enlarge all return statements? But you don't set the actual value? So why allocate it with 2 ops in the first place?? [Seems I completely missed that MPX changes "gimple" and the design document that was posted somewhere??]
I'm terribly concerned about the enlarging of the return statements. On the other hand, if something enlarged a GIMPLE_ASSIGN, then I would have called it out for further discussion.
Bah. Where is the update to gimple.texi and tree.texi?
Good catch. Ilya, please send a patch to update the docs.
