[PATCH] fortran: Release symbols in reversed order [PR106050]

[PATCH] fortran: Release symbols in reversed order [PR106050]

[Mikael Morin]

Hello,

I saw the light regarding this PR after Paul posted a comment yesterday.

Regression test in progress on x86_64-pc-linux-gnu.
I plan to push in the next hours.

Mikael

-- >8 --

Release symbols in reversed order wrt the order they were allocated.
This fixes an error recovery ICE in the case of a misplaced
derived type declaration.  Such a declaration creates nested
symbols, one for the derived type and one for each type parameter,
which should be immediately released as the declaration is
rejected.  This breaks if the derived type is released first.
As the type parameter symbols are in the namespace of the derived
type, releasing the derived type releases the type parameters, so
one can't access them after that, even to release them.  Hence,
the type parameters should be released first.

	PR fortran/106050

gcc/fortran/ChangeLog:

	* symbol.cc (gfc_restore_last_undo_checkpoint): Release symbols
	in reverse order.

gcc/testsuite/ChangeLog:

	* gfortran.dg/pdt_33.f90: New test.
---
 gcc/fortran/symbol.cc                |  2 +-
 gcc/testsuite/gfortran.dg/pdt_33.f90 | 15 +++++++++++++++
 2 files changed, 16 insertions(+), 1 deletion(-)
 create mode 100644 gcc/testsuite/gfortran.dg/pdt_33.f90

diff --git a/gcc/fortran/symbol.cc b/gcc/fortran/symbol.cc
index 37a9e8fa0ae..4a71d84b3fe 100644
--- a/gcc/fortran/symbol.cc
+++ b/gcc/fortran/symbol.cc
@@ -3661,7 +3661,7 @@ gfc_restore_last_undo_checkpoint (void)
   gfc_symbol *p;
   unsigned i;
 
-  FOR_EACH_VEC_ELT (latest_undo_chgset->syms, i, p)
+  FOR_EACH_VEC_ELT_REVERSE (latest_undo_chgset->syms, i, p)
     {
       /* Symbol in a common block was new. Or was old and just put in common */
       if (p->common_block
diff --git a/gcc/testsuite/gfortran.dg/pdt_33.f90 b/gcc/testsuite/gfortran.dg/pdt_33.f90
new file mode 100644
index 00000000000..0521513f2f8
--- /dev/null
+++ b/gcc/testsuite/gfortran.dg/pdt_33.f90
@@ -0,0 +1,15 @@
+! { dg-do compile }
+!
+! PR fortran/106050
+! The following used to trigger an error recovery ICE by releasing
+! the symbol T before the symbol K which was leading to releasing
+! K twice as it's in T's namespace.
+!
+! Contributed by G. Steinmetz <gscfq@t-online.de>
+
+program p
+   a = 1
+   type t(k)                  ! { dg-error "Unexpected derived type declaration" }
+      integer, kind :: k = 4  ! { dg-error "not allowed outside a TYPE definition" }
+   end type                   ! { dg-error "Expecting END PROGRAM" }
+end
-- 
2.40.1

Re: [PATCH] fortran: Release symbols in reversed order [PR106050]

[Paul Richard Thomas]

Hi Mikhail,

That's more than OK by me.

Thanks for attacking this PR.

I have a couple more of Steve's orphans waiting to be packaged up -
91960 and 104649. I'll submit them this evening.100607 is closed-fixed
and 103796 seems to be fixed.

Regards

Paul

On Tue, 11 Jul 2023 at 13:08, Mikael Morin via Fortran
<fortran@gcc.gnu.org> wrote:
>
> Hello,
>
> I saw the light regarding this PR after Paul posted a comment yesterday.
>
> Regression test in progress on x86_64-pc-linux-gnu.
> I plan to push in the next hours.
>
> Mikael
>
> -- >8 --
>
> Release symbols in reversed order wrt the order they were allocated.
> This fixes an error recovery ICE in the case of a misplaced
> derived type declaration.  Such a declaration creates nested
> symbols, one for the derived type and one for each type parameter,
> which should be immediately released as the declaration is
> rejected.  This breaks if the derived type is released first.
> As the type parameter symbols are in the namespace of the derived
> type, releasing the derived type releases the type parameters, so
> one can't access them after that, even to release them.  Hence,
> the type parameters should be released first.
>
>         PR fortran/106050
>
> gcc/fortran/ChangeLog:
>
>         * symbol.cc (gfc_restore_last_undo_checkpoint): Release symbols
>         in reverse order.
>
> gcc/testsuite/ChangeLog:
>
>         * gfortran.dg/pdt_33.f90: New test.
> ---
>  gcc/fortran/symbol.cc                |  2 +-
>  gcc/testsuite/gfortran.dg/pdt_33.f90 | 15 +++++++++++++++
>  2 files changed, 16 insertions(+), 1 deletion(-)
>  create mode 100644 gcc/testsuite/gfortran.dg/pdt_33.f90
>
> diff --git a/gcc/fortran/symbol.cc b/gcc/fortran/symbol.cc
> index 37a9e8fa0ae..4a71d84b3fe 100644
> --- a/gcc/fortran/symbol.cc
> +++ b/gcc/fortran/symbol.cc
> @@ -3661,7 +3661,7 @@ gfc_restore_last_undo_checkpoint (void)
>    gfc_symbol *p;
>    unsigned i;
>
> -  FOR_EACH_VEC_ELT (latest_undo_chgset->syms, i, p)
> +  FOR_EACH_VEC_ELT_REVERSE (latest_undo_chgset->syms, i, p)
>      {
>        /* Symbol in a common block was new. Or was old and just put in common */
>        if (p->common_block
> diff --git a/gcc/testsuite/gfortran.dg/pdt_33.f90 b/gcc/testsuite/gfortran.dg/pdt_33.f90
> new file mode 100644
> index 00000000000..0521513f2f8
> --- /dev/null
> +++ b/gcc/testsuite/gfortran.dg/pdt_33.f90
> @@ -0,0 +1,15 @@
> +! { dg-do compile }
> +!
> +! PR fortran/106050
> +! The following used to trigger an error recovery ICE by releasing
> +! the symbol T before the symbol K which was leading to releasing
> +! K twice as it's in T's namespace.
> +!
> +! Contributed by G. Steinmetz <gscfq@t-online.de>
> +
> +program p
> +   a = 1
> +   type t(k)                  ! { dg-error "Unexpected derived type declaration" }
> +      integer, kind :: k = 4  ! { dg-error "not allowed outside a TYPE definition" }
> +   end type                   ! { dg-error "Expecting END PROGRAM" }
> +end
> --
> 2.40.1
>


--
"If you can't explain it simply, you don't understand it well enough"
- Albert Einstein