public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Jason Merrill <jason@redhat.com>
To: Patrick Palka <ppalka@redhat.com>, gcc-patches@gcc.gnu.org
Subject: Re: [PATCH] c++: extract_local_specs and unevaluated contexts [PR100295]
Date: Thu, 15 Dec 2022 11:57:38 -0500	[thread overview]
Message-ID: <dfce9587-e4ac-df69-ce97-6196064d8e96@redhat.com> (raw)
In-Reply-To: <20221209215720.3142097-1-ppalka@redhat.com>

On 12/9/22 16:57, Patrick Palka wrote:
> Here during partial instantiation of the constexpr if, extra_local_specs
> walks the statement looking for local specializations within to save and
> possibly capture.  However, we're thwarted by the fact that 'ts' first
> appears inside an unevaluated context, and so the calls to
> process_outer_var_ref for its local specializations are a no-op.  And
> since we walk each tree exactly once, we end up not capturing them
> despite it later occuring in an evaluated context.
> 
> This patch fixes this by making extract_local_specs walk evaluated
> contexts first before walking unevaluated contexts.  We could probably
> get away with not walking unevaluated contexts at all, but this approach
> seems safer.
> 
> Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK for
> trunk/12?

OK.

> 	PR c++/100295
> 	PR c++/107579
> 
> gcc/cp/ChangeLog:
> 
> 	* pt.cc (el_data::skip_unevaluated_operands): New data member.
> 	(extract_locals_r): If skip_unevaluated_operands is true,
> 	don't walk into unevaluated contexts.
> 	(extract_local_specs): Walk the pattern twice, first with
> 	skip_unevaluated_operands true followed by it set to false.
> 
> gcc/testsuite/ChangeLog:
> 
> 	* g++.dg/cpp1z/constexpr-if-lambda5.C: New test.
> ---
>   gcc/cp/pt.cc                                  | 19 ++++++++++++++++++-
>   .../g++.dg/cpp1z/constexpr-if-lambda5.C       | 15 +++++++++++++++
>   2 files changed, 33 insertions(+), 1 deletion(-)
>   create mode 100644 gcc/testsuite/g++.dg/cpp1z/constexpr-if-lambda5.C
> 
> diff --git a/gcc/cp/pt.cc b/gcc/cp/pt.cc
> index d05a49b1c11..2b22bf14c53 100644
> --- a/gcc/cp/pt.cc
> +++ b/gcc/cp/pt.cc
> @@ -13015,17 +13015,26 @@ public:
>     /* List of local_specializations used within the pattern.  */
>     tree extra;
>     tsubst_flags_t complain;
> +  /* True iff we don't want to walk into unevaluated contexts.  */
> +  bool skip_unevaluated_operands = false;
>   
>     el_data (tsubst_flags_t c)
>       : extra (NULL_TREE), complain (c) {}
>   };
>   static tree
> -extract_locals_r (tree *tp, int */*walk_subtrees*/, void *data_)
> +extract_locals_r (tree *tp, int *walk_subtrees, void *data_)
>   {
>     el_data &data = *reinterpret_cast<el_data*>(data_);
>     tree *extra = &data.extra;
>     tsubst_flags_t complain = data.complain;
>   
> +  if (data.skip_unevaluated_operands
> +      && unevaluated_p (TREE_CODE (*tp)))
> +    {
> +      *walk_subtrees = 0;
> +      return NULL_TREE;
> +    }
> +
>     if (TYPE_P (*tp) && typedef_variant_p (*tp))
>       /* Remember local typedefs (85214).  */
>       tp = &TYPE_NAME (*tp);
> @@ -13117,6 +13126,14 @@ static tree
>   extract_local_specs (tree pattern, tsubst_flags_t complain)
>   {
>     el_data data (complain);
> +  /* Walk the pattern twice, ignoring unevaluated operands the first time
> +     around, so that if a local specialization appears in both an
> +     evaluated and unevaluated context we prefer to process it in the
> +     former context (since e.g. process_outer_var_ref is a no-op inside
> +     an unevaluated context).  */
> +  data.skip_unevaluated_operands = true;
> +  cp_walk_tree (&pattern, extract_locals_r, &data, &data.visited);
> +  data.skip_unevaluated_operands = false;
>     cp_walk_tree (&pattern, extract_locals_r, &data, &data.visited);
>     return data.extra;
>   }
> diff --git a/gcc/testsuite/g++.dg/cpp1z/constexpr-if-lambda5.C b/gcc/testsuite/g++.dg/cpp1z/constexpr-if-lambda5.C
> new file mode 100644
> index 00000000000..d2bf0221743
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/cpp1z/constexpr-if-lambda5.C
> @@ -0,0 +1,15 @@
> +// PR c++/100295
> +// { dg-do compile { target c++17 } }
> +
> +template<typename... Ts>
> +void f(Ts... ts) {
> +  auto lambda = [=](auto x) {
> +    if constexpr (sizeof((ts+x) + ...) != 0)
> +      (..., ts);
> +  };
> +  lambda(0);
> +}
> +
> +int main() {
> +  f(0, 'a');
> +}


      reply	other threads:[~2022-12-15 16:57 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-12-09 21:57 Patrick Palka
2022-12-15 16:57 ` Jason Merrill [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=dfce9587-e4ac-df69-ce97-6196064d8e96@redhat.com \
    --to=jason@redhat.com \
    --cc=gcc-patches@gcc.gnu.org \
    --cc=ppalka@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).