On Thu, Nov 05 2015, Bernd Schmidt wrote:
> On 11/05/2015 12:33 PM, Andreas Arnez wrote:
>
>> Thanks again for reviewing. Are you going to look at patch #2 as well?
>
> Yeah, still thinking about that one.
>
>>> Does C++ have similar issues?
>>
>> Not this particular issue, AFAIK. But I've just looked at how C++ fares
>> with the enhanced version of pr67192.c from patch #2. There I see the
>> following:
>>
>> Breakpoint 2, f4 () at pr67192.cc:54
>> (gdb) p cnt
>> $1 = 16
>>
>> I.e., when breaking on "while (1)" the first loop iteration has already
>> executed. This is because the C++ parser assigns the backward-goto to
>> the 'while' token. It's the same issue you pointed at with version 2 of
>> my patch.
>>
>> Shall I open a bug for that?
>
> I'd obviously prefer if you'd manage to get the two frontends behave
> identically. The alternative would be to open a bug.
OK, I guess it depends on whether we want to go the route of patch #2.
If so, it seems we can do the same for the C++ parser, like in the patch
below. Note that I've not tested this very much.
-- >8 --
Subject: [PATCH] C++: Fix location of loop statement
---
gcc/cp/cp-gimplify.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/gcc/cp/cp-gimplify.c b/gcc/cp/cp-gimplify.c
index e4b50e5..d9bb708 100644
--- a/gcc/cp/cp-gimplify.c
+++ b/gcc/cp/cp-gimplify.c
@@ -266,7 +266,12 @@ genericize_cp_loop (tree *stmt_p, location_t start_locus,
tree cond, tree body,
loop = stmt_list;
}
else
- loop = build1_loc (start_locus, LOOP_EXPR, void_type_node, stmt_list);
+ {
+ location_t loc = EXPR_LOCATION (expr_first (body));
+ if (loc == UNKNOWN_LOCATION)
+ loc = start_locus;
+ loop = build1_loc (loc, LOOP_EXPR, void_type_node, stmt_list);
+ }
stmt_list = NULL;
append_to_statement_list (loop, &stmt_list);
--
2.3.0
