public inbox for fortran@gcc.gnu.org
 help / color / mirror / Atom feed
* [Patch, fortran] PR82550 - program using submodules fails to link
@ 2017-10-17 18:33 Paul Richard Thomas
  2017-10-18  0:52 ` Jerry DeLisle
  0 siblings, 1 reply; 3+ messages in thread
From: Paul Richard Thomas @ 2017-10-17 18:33 UTC (permalink / raw)
  To: fortran, gcc-patches

[-- Attachment #1: Type: text/plain, Size: 603 bytes --]

The attached patch has a comment that explains what is going on.

Bootstrapped and regtested on FC23/x86_64 - OK for trunk and 7-branch?

Paul

2017-10-17  Paul Thomas  <pault@gcc.gnu.org>

    PR fortran/82550
    * expr.c (gfc_check_pointer_assign): A use associated procedure
    target in a submodule must have the 'use_assoc' attribute set
    so that the name mangling is done correctly.

2017-10-17  Paul Thomas  <pault@gcc.gnu.org>

    PR fortran/82550
    * gfortran.dg/submodule_30.f08 : New test.


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

[-- Attachment #2: submit.diff --]
[-- Type: text/plain, Size: 2018 bytes --]

Index: gcc/fortran/expr.c
===================================================================
*** gcc/fortran/expr.c	(revision 253748)
--- gcc/fortran/expr.c	(working copy)
*************** gfc_check_pointer_assign (gfc_expr *lval
*** 3632,3637 ****
--- 3632,3645 ----
  	  name = s2->name;
  	}
  
+       /* Make the procedure use associated so that the middle end does
+ 	 the right thing with name mangling. This undoes the reset in
+ 	 parse.c(set_syms_host_assoc) and is necessary to allow the
+ 	 attributes of module procedure interfaces to be changed.  */
+       if (s2 && s2->attr.flavor == FL_PROCEDURE
+ 	  && s2->module && s2->attr.used_in_submodule)
+ 	s2->attr.use_assoc = 1;
+ 
        if (s2 && s2->attr.proc_pointer && s2->ts.interface)
  	s2 = s2->ts.interface;
  
Index: gcc/testsuite/gfortran.dg/submodule_30.f08
===================================================================
*** gcc/testsuite/gfortran.dg/submodule_30.f08	(nonexistent)
--- gcc/testsuite/gfortran.dg/submodule_30.f08	(working copy)
***************
*** 0 ****
--- 1,42 ----
+ ! { dg-do run }
+ !
+ ! Test the fix for PR82550 in which the reference to 'p' in 'foo'
+ ! was not being correctly handled.
+ !
+ ! Contributed by Reinhold Bader  <Bader@lrz.de>
+ !
+ module m_subm_18_pos
+   implicit none
+   integer :: i = 0
+   interface
+     module subroutine foo(fun_ptr)
+       procedure(p), pointer, intent(out) :: fun_ptr
+     end subroutine
+   end interface
+ contains
+   subroutine p()
+     i = 1
+   end subroutine p
+ end module m_subm_18_pos
+ submodule (m_subm_18_pos) subm_18_pos
+     implicit none
+ contains
+     module subroutine foo(fun_ptr)
+       procedure(p), pointer, intent(out) :: fun_ptr
+       fun_ptr => p
+     end subroutine
+ end submodule
+ program p_18_pos
+   use m_subm_18_pos
+   implicit none
+   procedure(), pointer :: x
+   call foo(x)
+   call x()
+   if (i == 1) then
+      write(*,*) 'OK'
+   else
+      write(*,*) 'FAIL'
+      call abort
+   end if
+ end program p_18_pos
+ 

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [Patch, fortran] PR82550 - program using submodules fails to link
  2017-10-17 18:33 [Patch, fortran] PR82550 - program using submodules fails to link Paul Richard Thomas
@ 2017-10-18  0:52 ` Jerry DeLisle
  2017-10-18  8:58   ` Paul Richard Thomas
  0 siblings, 1 reply; 3+ messages in thread
From: Jerry DeLisle @ 2017-10-18  0:52 UTC (permalink / raw)
  To: Paul Richard Thomas, fortran, gcc-patches

On 10/17/2017 11:33 AM, Paul Richard Thomas wrote:
> The attached patch has a comment that explains what is going on.
> 
> Bootstrapped and regtested on FC23/x86_64 - OK for trunk and 7-branch?
> 

Yes, looks OK for both. Thanks.

Jerry

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [Patch, fortran] PR82550 - program using submodules fails to link
  2017-10-18  0:52 ` Jerry DeLisle
@ 2017-10-18  8:58   ` Paul Richard Thomas
  0 siblings, 0 replies; 3+ messages in thread
From: Paul Richard Thomas @ 2017-10-18  8:58 UTC (permalink / raw)
  To: Jerry DeLisle; +Cc: fortran, gcc-patches

[-- Attachment #1: Type: text/plain, Size: 934 bytes --]

Dear Jerry,

Thanks. I Committed as revision 253848 a still simpler patch that
achieves the same end. Please see the attached and the ChangeLogs
below.

Regards

Paul

2017-10-18  Paul Thomas  <pault@gcc.gnu.org>

    PR fortran/82550
    * trans_decl.c (gfc_get_symbol_decl): Procedure symbols that
    have the 'used_in_submodule' attribute should be processed by
    'gfc_get_extern_function_decl'.

2017-10-18  Paul Thomas  <pault@gcc.gnu.org>

    PR fortran/82550
    * gfortran.dg/submodule_30.f08 : New test.

On 18 October 2017 at 01:52, Jerry DeLisle <jvdelisle@charter.net> wrote:
> On 10/17/2017 11:33 AM, Paul Richard Thomas wrote:
>> The attached patch has a comment that explains what is going on.
>>
>> Bootstrapped and regtested on FC23/x86_64 - OK for trunk and 7-branch?
>>
>
> Yes, looks OK for both. Thanks.
>
> Jerry



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

[-- Attachment #2: submit.diff --]
[-- Type: text/plain, Size: 2257 bytes --]

Index: /home/pault/svn/trunk/gcc/fortran/trans-decl.c
===================================================================
*** /home/pault/svn/trunk/gcc/fortran/trans-decl.c	(revision 253748)
--- /home/pault/svn/trunk/gcc/fortran/trans-decl.c	(working copy)
*************** gfc_get_symbol_decl (gfc_symbol * sym)
*** 1670,1676 ****
      {
        /* Catch functions. Only used for actual parameters,
  	 procedure pointers and procptr initialization targets.  */
!       if (sym->attr.use_assoc || sym->attr.intrinsic
  	  || sym->attr.if_source != IFSRC_DECL)
  	{
  	  decl = gfc_get_extern_function_decl (sym);
--- 1670,1678 ----
      {
        /* Catch functions. Only used for actual parameters,
  	 procedure pointers and procptr initialization targets.  */
!       if (sym->attr.use_assoc
! 	  || sym->attr.used_in_submodule
! 	  || sym->attr.intrinsic
  	  || sym->attr.if_source != IFSRC_DECL)
  	{
  	  decl = gfc_get_extern_function_decl (sym);
Index: /home/pault/svn/trunk/gcc/testsuite/gfortran.dg/submodule_30.f08
===================================================================
*** /home/pault/svn/trunk/gcc/testsuite/gfortran.dg/submodule_30.f08	(nonexistent)
--- /home/pault/svn/trunk/gcc/testsuite/gfortran.dg/submodule_30.f08	(working copy)
***************
*** 0 ****
--- 1,42 ----
+ ! { dg-do run }
+ !
+ ! Test the fix for PR82550 in which the reference to 'p' in 'foo'
+ ! was not being correctly handled.
+ !
+ ! Contributed by Reinhold Bader  <Bader@lrz.de>
+ !
+ module m_subm_18_pos
+   implicit none
+   integer :: i = 0
+   interface
+     module subroutine foo(fun_ptr)
+       procedure(p), pointer, intent(out) :: fun_ptr
+     end subroutine
+   end interface
+ contains
+   subroutine p()
+     i = 1
+   end subroutine p
+ end module m_subm_18_pos
+ submodule (m_subm_18_pos) subm_18_pos
+     implicit none
+ contains
+     module subroutine foo(fun_ptr)
+       procedure(p), pointer, intent(out) :: fun_ptr
+       fun_ptr => p
+     end subroutine
+ end submodule
+ program p_18_pos
+   use m_subm_18_pos
+   implicit none
+   procedure(), pointer :: x
+   call foo(x)
+   call x()
+   if (i == 1) then
+      write(*,*) 'OK'
+   else
+      write(*,*) 'FAIL'
+      call abort
+   end if
+ end program p_18_pos
+ 

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2017-10-18  8:58 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-10-17 18:33 [Patch, fortran] PR82550 - program using submodules fails to link Paul Richard Thomas
2017-10-18  0:52 ` Jerry DeLisle
2017-10-18  8:58   ` Paul Richard Thomas

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).