Re: [PATCH] PR c++/68795: fix uninitialized close_paren_loc in cp_parser_postfix_expression (v2)

[Bernd Schmidt]

On 12/18/2015 08:05 PM, David Malcolm wrote:
On Thu, 2015-12-17 at 19:21 +0100, Bernd Schmidt wrote:
On 12/17/2015 07:32 PM, David Malcolm wrote:
+             if (close_paren_loc)

close_paren_loc != UNKNOWN_LOCATION - it's very confusing otherwise.


Bernd

Here's an updated version; the only change since v1 is:
-           if (close_paren_loc)
+           if (close_paren_loc != UNKNOWN_LOCATION)

Have verified the fix in valgrind.  OK for trunk if it still passes
bootstrap&regrtest?

gcc/cp/ChangeLog:
        * parser.c (cp_parser_postfix_expression): Initialize
        close_paren_loc to UNKNOWN_LOCATION; only use it if
        it has been written to by
        cp_parser_parenthesized_expression_list.
        (cp_parser_postfix_dot_deref_expression): Likewise.
        (cp_parser_parenthesized_expression_list): Document the behavior
        with respect to the CLOSE_PAREN_LOC param.

I usually like to leave C++ patches to Jason, but I guess I don't need 
to know the standard inside out for this one.

Prior to your ealier patch, we had

protected_set_expr_location (postfix_expression, token->location);

It looks like after your new patch, we could end up not setting the 
location on the expr at all if close_paren_loc is still 
UNKNOWN_LOCATION. I'm guessing this is intentional as the comment update 
suggests this is an error path, and the cp_expr constructor ensures that 
we get at least UNKNOWN_LOCATION and not another uninitialized loc. If 
I'm correct in all that, the patch is ok.


Bernd