Before the fix to the ranger dependency chain yesterday (commit
7f0cfeb1) I thought an ICE I was seeing was due to my get_range_query
patchet. This was not the case, but this small change crept in while I
was debugging the failure.
I'm reverting the change as was approved.
Tested on x86-64 Linux.
Pushed.
gcc/ChangeLog:
* vr-values.c (simplify_conversion_using_ranges): Use
get_range_query instead of get_global_range_query.
---
gcc/vr-values.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/gcc/vr-values.c b/gcc/vr-values.c
index d283108b7c2..3d0be8edb3b 100644
--- a/gcc/vr-values.c
+++ b/gcc/vr-values.c
@@ -3837,7 +3837,7 @@ simplify_conversion_using_ranges (gimple_stmt_iterator
*gsi, gimple *stmt)
value_range vr;
if (!INTEGRAL_TYPE_P (TREE_TYPE (innerop)))
return false;
- get_global_range_query ()->range_of_expr (vr, innerop, stmt);
+ get_range_query (cfun)->range_of_expr (vr, innerop, stmt);
if (vr.undefined_p () || vr.varying_p ())
return false;
innermin = widest_int::from (vr.lower_bound (), TYPE_SIGN (TREE_TYPE
(innerop)));
--
2.31.1
