[PATCH] Fortran: avoid obsolescence warning for COMMON with submodule [PR111880]

[PATCH] Fortran: avoid obsolescence warning for COMMON with submodule [PR111880]

[Harald Anlauf]

[-- Attachment #1: Type: text/plain, Size: 491 bytes --]

Dear all,

the PR is about a redundant obsolescence warning for COMMON when
a symbols appears in the scope of a submodule.  As we did not warn
for use-associated symbols, it seemed natural to extend this to
symbols that are used in a submodule.  Or am I missing anything?

Regtests cleanly on x86_64-pc-linux-gnu.  OK for mainline?

The PR is marked as a regression (the warning appeared in gcc-9).
It looks simple enough for backporting, or does anybody see any
risk here?

Thanks,
Harald


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: pr111880.diff --]
[-- Type: text/x-patch, Size: 2316 bytes --]

From a962ab0417f5ff2efd51e710ae370d9f4a4b9f1a Mon Sep 17 00:00:00 2001
From: Harald Anlauf <anlauf@gmx.de>
Date: Thu, 23 Nov 2023 22:48:38 +0100
Subject: [PATCH] Fortran: avoid obsolescence warning for COMMON with submodule
 [PR111880]

gcc/fortran/ChangeLog:

	PR fortran/111880
	* resolve.cc (resolve_common_vars): Do not call gfc_add_in_common
	for symbols that are USE associated or used in a submodule.

gcc/testsuite/ChangeLog:

	PR fortran/111880
	* gfortran.dg/pr111880.f90: New test.
---
 gcc/fortran/resolve.cc                 |  4 ++--
 gcc/testsuite/gfortran.dg/pr111880.f90 | 22 ++++++++++++++++++++++
 2 files changed, 24 insertions(+), 2 deletions(-)
 create mode 100644 gcc/testsuite/gfortran.dg/pr111880.f90

diff --git a/gcc/fortran/resolve.cc b/gcc/fortran/resolve.cc
index 81a14653a04..166b702cd9a 100644
--- a/gcc/fortran/resolve.cc
+++ b/gcc/fortran/resolve.cc
@@ -986,8 +986,8 @@ resolve_common_vars (gfc_common_head *common_block, bool named_common)

       /* gfc_add_in_common may have been called before, but the reported errors
 	 have been ignored to continue parsing.
-	 We do the checks again here.  */
-      if (!csym->attr.use_assoc)
+	 We do the checks again here, unless the symbol is USE associated.  */
+      if (!csym->attr.use_assoc && !csym->attr.used_in_submodule)
 	{
 	  gfc_add_in_common (&csym->attr, csym->name, &common_block->where);
 	  gfc_notify_std (GFC_STD_F2018_OBS, "COMMON block at %L",
diff --git a/gcc/testsuite/gfortran.dg/pr111880.f90 b/gcc/testsuite/gfortran.dg/pr111880.f90
new file mode 100644
index 00000000000..c0cd98a93d4
--- /dev/null
+++ b/gcc/testsuite/gfortran.dg/pr111880.f90
@@ -0,0 +1,22 @@
+! { dg-do compile }
+! { dg-options "-std=f2018" }
+! PR fortran/111880 - redundant warning of obsolescent COMMON with submodule
+
+module third_party_module
+  integer :: some_param
+  common /not_my_code/ some_param   ! { dg-warning "COMMON block" }
+end module third_party_module
+
+module foo
+  use third_party_module
+  interface
+    module subroutine bar()
+    end subroutine bar
+  end interface
+end module foo
+
+submodule (foo) foo_submod      ! We do not need a warning here!
+contains
+  module procedure bar
+  end procedure bar
+end submodule foo_submod
--
2.35.3

Re: [PATCH] Fortran: avoid obsolescence warning for COMMON with submodule [PR111880]

[Mikael Morin]

Hello,

Le 23/11/2023 à 22:59, Harald Anlauf a écrit :
> Dear all,
> 
> the PR is about a redundant obsolescence warning for COMMON when
> a symbols appears in the scope of a submodule.  As we did not warn
> for use-associated symbols, it seemed natural to extend this to
> symbols that are used in a submodule.  Or am I missing anything?
> 
> Regtests cleanly on x86_64-pc-linux-gnu.  OK for mainline?
> 
Yes, looks good.  Obvious I'm tempted to say.

> The PR is marked as a regression (the warning appeared in gcc-9).
> It looks simple enough for backporting, or does anybody see any
> risk here?
> 
No, you can go ahead.

Thanks
Mikael