On 12/15/21 09:09, Jakub Jelinek wrote:
On Wed, Dec 15, 2021 at 08:58:45AM -0500, Patrick Palka wrote:
Oops, thanks Jakub, I didn't realize we don't run the testsuite with
-std=c++23 yet.
I guess it'd be too expensive to add another std to the testing matrix
at this point, but I wonder if the
On Wed, Dec 15, 2021 at 08:58:45AM -0500, Patrick Palka wrote:
> Oops, thanks Jakub, I didn't realize we don't run the testsuite with
> -std=c++23 yet.
>
> I guess it'd be too expensive to add another std to the testing matrix
> at this point, but I wonder if the test harness should at least run
On Wed, 15 Dec 2021, Jakub Jelinek wrote:
> On Mon, Dec 13, 2021 at 04:28:26PM -0500, Patrick Palka via Gcc-patches wrote:
> > * g++.dg/concepts/diagnostic18.C: Expect a "constraints on a
> > non-templated function" error.
> > * g++.dg/cpp23/auto-fncast10.C: New test.
>
> This test
On Mon, Dec 13, 2021 at 04:28:26PM -0500, Patrick Palka via Gcc-patches wrote:
> * g++.dg/concepts/diagnostic18.C: Expect a "constraints on a
> non-templated function" error.
> * g++.dg/cpp23/auto-fncast10.C: New test.
This test fails:
+FAIL: g++.dg/cpp23/auto-fncast11.C
as
processing_template_decl, or use current_template_parms directly. OK
whichever way you prefer.
-- >8 --
Subject: [PATCH] c++: processing_template_decl vs template depth [PR103408]
We use processing_template_decl in two slightly different ways: as a
flag to signal that we're dealing with templated tr
working the calculation of
friend_depth in make_friend_class.
Also replace processing_template_decl with
current_template_decl in PROCESSING_REAL_TEMPLATE_DECL_P,
start_preparsed_function and push_template_decl. These were
missed in the original patch because I hadn't considered grepping
On 12/10/21 14:12, Patrick Palka wrote:
We use processing_template_decl in two slightly different ways: as
a flag to signal that we're dealing with templated trees, and its
magnitude is also used as a proxy for the current syntactic template
depth. This overloaded meaning of p_t_d is
We use processing_template_decl in two slightly different ways: as
a flag to signal that we're dealing with templated trees, and its
magnitude is also used as a proxy for the current syntactic template
depth. This overloaded meaning of p_t_d is conceptually confusing and
leads to bugs that we end
