[PATCH] [PR86823] retain deferred access checks from outside firewall

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

[PATCH] [PR86823] retain deferred access checks from outside firewall

Alexandre Oliva-2
We used to preserve deferred access check along with resolved template
ids, but a tentative parsing firewall introduced additional layers of
deferred access checks, so that we don't preserve the checks we
want to any more.

This patch collapses the access check levels introduced by the
firewall with those we wanted to preserve to begin with, saves them,
pushes them one more layer up so that the firewall doesn't drop them,
and then pushes new empty levels so balance out with the upcoming
popping.

We may want to avoid the pop_to_parent and one push after the save, and
instead leave the firewall's scope before the final pop_to_parent.
However, this patch is what I regression-tested successfully with
check-g++ on x86_64-linux-gnu.  Regstrapping on i686- and
x86_64-linux-gnu now.  Ok to install?


for  gcc/cp/ChangeLog

        PR c++/86823
        * parser.c (cp_parser_template_id): Merge access checks from
        outside the tentative parsing firewall with those from inside,
        and save them all with the template id token.

for  gcc/testsuite/ChangeLog

        PR c++/86823
        * g++.dg/pr86823.C: New.
---
 gcc/cp/parser.c                |   12 ++++++++++++
 gcc/testsuite/g++.dg/pr86823.C |   15 +++++++++++++++
 2 files changed, 27 insertions(+)
 create mode 100644 gcc/testsuite/g++.dg/pr86823.C

diff --git a/gcc/cp/parser.c b/gcc/cp/parser.c
index db0f0338179e..6a08b09715a7 100644
--- a/gcc/cp/parser.c
+++ b/gcc/cp/parser.c
@@ -16185,7 +16185,19 @@ cp_parser_template_id (cp_parser *parser,
  so the memory will not be reclaimed during token replacing below.  */
       token->u.tree_check_value = ggc_cleared_alloc<struct tree_check> ();
       token->u.tree_check_value->value = template_id;
+      /* Collapse the access check levels introduced by
+ tentative_firewall with the one we introduced ourselves.  */
+      pop_to_parent_deferring_access_checks ();
+      pop_to_parent_deferring_access_checks ();
       token->u.tree_check_value->checks = get_deferred_access_checks ();
+      /* Push the checks up one more level, so that the firewall
+ doesn't drop them on the floor when we return.  Then, push
+ empty levels back in place so that they are popped
+ properly.  */
+      pop_to_parent_deferring_access_checks ();
+      push_deferring_access_checks (dk_deferred);
+      push_deferring_access_checks (dk_deferred);
+      push_deferring_access_checks (dk_deferred);
       token->keyword = RID_MAX;
 
       /* Purge all subsequent tokens.  */
diff --git a/gcc/testsuite/g++.dg/pr86823.C b/gcc/testsuite/g++.dg/pr86823.C
new file mode 100644
index 000000000000..18914b00aa8d
--- /dev/null
+++ b/gcc/testsuite/g++.dg/pr86823.C
@@ -0,0 +1,15 @@
+// { dg-do compile }
+
+struct X {
+private:
+  template<typename T>
+  struct Y {
+    int data;
+  };
+public:
+  int value;
+};
+
+int main() {
+  typename X::Y<int> a; // { dg-error "private" }
+}

--
Alexandre Oliva, freedom fighter   https://FSFLA.org/blogs/lxo
Be the change, be Free!         FSF Latin America board member
GNU Toolchain Engineer                Free Software Evangelist
Hay que enGNUrecerse, pero sin perder la terGNUra jamás-GNUChe
Reply | Threaded
Open this post in threaded view
|

Re: [PATCH] [PR86823] retain deferred access checks from outside firewall

Jason Merrill
On 11/16/18 9:16 PM, Alexandre Oliva wrote:

> We used to preserve deferred access check along with resolved template
> ids, but a tentative parsing firewall introduced additional layers of
> deferred access checks, so that we don't preserve the checks we
> want to any more.
>
> This patch collapses the access check levels introduced by the
> firewall with those we wanted to preserve to begin with, saves them,
> pushes them one more layer up so that the firewall doesn't drop them,
> and then pushes new empty levels so balance out with the upcoming
> popping.
>
> We may want to avoid the pop_to_parent and one push after the save, and
> instead leave the firewall's scope before the final pop_to_parent.
> However, this patch is what I regression-tested successfully with
> check-g++ on x86_64-linux-gnu.  Regstrapping on i686- and
> x86_64-linux-gnu now.  Ok to install?
>
>
> for  gcc/cp/ChangeLog
>
> PR c++/86823
> * parser.c (cp_parser_template_id): Merge access checks from
> outside the tentative parsing firewall with those from inside,
> and save them all with the template id token.
>
> for  gcc/testsuite/ChangeLog
>
> PR c++/86823
> * g++.dg/pr86823.C: New.
> ---
>   gcc/cp/parser.c                |   12 ++++++++++++
>   gcc/testsuite/g++.dg/pr86823.C |   15 +++++++++++++++
>   2 files changed, 27 insertions(+)
>   create mode 100644 gcc/testsuite/g++.dg/pr86823.C
>
> diff --git a/gcc/cp/parser.c b/gcc/cp/parser.c
> index db0f0338179e..6a08b09715a7 100644
> --- a/gcc/cp/parser.c
> +++ b/gcc/cp/parser.c
> @@ -16185,7 +16185,19 @@ cp_parser_template_id (cp_parser *parser,
>   so the memory will not be reclaimed during token replacing below.  */
>         token->u.tree_check_value = ggc_cleared_alloc<struct tree_check> ();
>         token->u.tree_check_value->value = template_id;
> +      /* Collapse the access check levels introduced by
> + tentative_firewall with the one we introduced ourselves.  */
> +      pop_to_parent_deferring_access_checks ();
> +      pop_to_parent_deferring_access_checks ();
>         token->u.tree_check_value->checks = get_deferred_access_checks ();
> +      /* Push the checks up one more level, so that the firewall
> + doesn't drop them on the floor when we return.  Then, push
> + empty levels back in place so that they are popped
> + properly.  */
> +      pop_to_parent_deferring_access_checks ();
> +      push_deferring_access_checks (dk_deferred);
> +      push_deferring_access_checks (dk_deferred);
> +      push_deferring_access_checks (dk_deferred);
Hmm, I'm uncomfortable with how this depends on the specific
implementation of tentative_firewall.  What do you think of this
alternate approach (untested other than with the testcase)?

Jason

checks.diff (1K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: [PATCH] [PR86823] retain deferred access checks from outside firewall

Alexandre Oliva-2
On Dec  5, 2018, Jason Merrill <[hidden email]> wrote:

> Hmm, I'm uncomfortable with how this depends on the specific
> implementation of tentative_firewall.

*nod*

> What do you think of this alternate approach (untested other than with
> the testcase)?

If that won't drop any other deferred access checks that we might want
to keep when leaving the firewall by other paths, it looks much nicer
indeed!  I had been careful to only move access checks when I was
certain about the replacement, but even there I was concerned about
carrying stuff along with the preparsed token that didn't belong.

I'm giving your proposed patch a full round of testing along with other
patches.

--
Alexandre Oliva, freedom fighter   https://FSFLA.org/blogs/lxo
Be the change, be Free!         FSF Latin America board member
GNU Toolchain Engineer                Free Software Evangelist
Hay que enGNUrecerse, pero sin perder la terGNUra jamás-GNUChe