[Bug c++/104379] New: [p/10/11/12 Regression] -Wshadow warning given three times

[Bug c++/104379] New: [p/10/11/12 Regression] -Wshadow warning given three times

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379

            Bug ID: 104379
           Summary: [p/10/11/12 Regression] -Wshadow warning given three
                    times
           Product: gcc
           Version: 11.2.1
            Status: UNCONFIRMED
          Keywords: diagnostic
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: redi at gcc dot gnu.org
  Target Milestone: ---

struct S {
  int x;
  S(int x) { (void)x; }
};


shad.C: In constructor 'S::S(int)':
shad.C:3:12: warning: declaration of 'x' shadows a member of 'S' [-Wshadow]
    3 |   S(int x) { (void)x; }
      |            ^
shad.C:2:7: note: shadowed declaration is here
    2 |   int x;
      |       ^
shad.C: In constructor 'S::S(int)':
shad.C:3:23: warning: declaration of 'x' shadows a member of 'S' [-Wshadow]
    3 |   S(int x) { (void)x; }
      |                       ^
shad.C:2:7: note: shadowed declaration is here
    2 |   int x;
      |       ^
shad.C: In constructor 'S::S(int)':
shad.C:3:23: warning: declaration of 'x' shadows a member of 'S' [-Wshadow]
    3 |   S(int x) { (void)x; }
      |                       ^
shad.C:2:7: note: shadowed declaration is here
    2 |   int x;
      |       ^

[Bug c++/104379] [9/10/11/12 Regression] -Wshadow warning given three times

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|                            |8.4.0
   Last reconfirmed|                            |2022-02-04
            Summary|[p/10/11/12 Regression]     |[9/10/11/12 Regression]
                   |-Wshadow warning given      |-Wshadow warning given
                   |three times                 |three times
             Status|UNCONFIRMED                 |NEW
      Known to fail|                            |10.3.0, 11.2.0, 12.0, 9.1.0
     Ever confirmed|0                           |1

[Bug c++/104379] [9/10/11/12 Regression] -Wshadow warning given three times

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |rguenth at gcc dot gnu.org

--- Comment #1 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Started with r264779

c-decl.c (warn_if_shadowing): Do not test DECL_FROM_INLINE.

[Bug c++/104379] [9/10/11/12 Regression] -Wshadow warning given three times

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |9.5

[Bug c++/104379] [9/10/11/12 Regression] -Wshadow warning given three times

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
I suspect we warn once for each CTOR clone, whilst with checking
DECL_FROM_INLINE
we excluded all but the master clone.  "from inline" is of course misleading
here.  I suppose the same issue might happen with templates where we'd warn
once per instantiation?

There's for sure a better way to exclude the CTOR/DTOR clones than checking
DECL_ABSTRACT_ORIGIN.

DECL_FROM_INLINE is nearly dead, it should probably be removed.

[Bug c++/104379] [9/10/11/12 Regression] -Wshadow warning given three times

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379

--- Comment #3 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to Richard Biener from comment #2)
> I suppose the same issue might happen with templates where we'd warn
> once per instantiation?

Yes indeed. Once for the primary template, and then again for each
specialization generated from it:


template<typename T>
struct S
{
  int i;
  int f(int i) { return i; }
};

int i = S<int>().f(1);
int j = S<long>().f(1);



shad2.C: In member function ‘int S<T>::f(int)’:
shad2.C:5:13: warning: declaration of ‘i’ shadows a member of ‘S<T>’ [-Wshadow]
    5 |   int f(int i) { return i; }
      |         ~~~~^
shad2.C:4:7: note: shadowed declaration is here
    4 |   int i;
      |       ^
shad2.C: In instantiation of ‘int S<T>::f(int) [with T = int]’:
shad2.C:8:19:   required from here
shad2.C:5:13: warning: declaration of ‘i’ shadows a member of ‘S<int>’
[-Wshadow]
    5 |   int f(int i) { return i; }
      |         ~~~~^
shad2.C:4:7: note: shadowed declaration is here
    4 |   int i;
      |       ^
shad2.C: In instantiation of ‘int S<T>::f(int) [with T = long int]’:
shad2.C:9:20:   required from here
shad2.C:5:13: warning: declaration of ‘i’ shadows a member of ‘S<long int>’
[-Wshadow]
    5 |   int f(int i) { return i; }
      |         ~~~~^
shad2.C:4:7: note: shadowed declaration is here
    4 |   int i;
      |       ^

[Bug c++/104379] [9/10/11/12 Regression] -Wshadow warning given three times

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379

--- Comment #4 from Jonathan Wakely <redi at gcc dot gnu.org> ---
So you can imagine what happens if you combine constructor clones with
templates! :-D

template<typename T>
struct S
{
  int i;
  S(int i) { (void) i; }
};

S<int> i(1);
S<long> j(1);

wheeeeeeeee!

shad2.C: In constructor ‘S<T>::S(int)’:
shad2.C:5:9: warning: declaration of ‘i’ shadows a member of ‘S<T>’ [-Wshadow]
    5 |   S(int i) { (void) i; }
      |     ~~~~^
shad2.C:4:7: note: shadowed declaration is here
    4 |   int i;
      |       ^
shad2.C: In instantiation of ‘S<T>::S(int) [with T = int]’:
shad2.C:8:11:   required from here
shad2.C:5:9: warning: declaration of ‘i’ shadows a member of ‘S<int>’
[-Wshadow]
    5 |   S(int i) { (void) i; }
      |     ~~~~^
shad2.C:4:7: note: shadowed declaration is here
    4 |   int i;
      |       ^
shad2.C:5:9: warning: declaration of ‘i’ shadows a member of ‘S<int>’
[-Wshadow]
    5 |   S(int i) { (void) i; }
      |     ~~~~^
shad2.C:4:7: note: shadowed declaration is here
    4 |   int i;
      |       ^
shad2.C:5:9: warning: declaration of ‘i’ shadows a member of ‘S<int>’
[-Wshadow]
    5 |   S(int i) { (void) i; }
      |     ~~~~^
shad2.C:4:7: note: shadowed declaration is here
    4 |   int i;
      |       ^
shad2.C: In instantiation of ‘S<T>::S(int) [with T = long int]’:
shad2.C:9:12:   required from here
shad2.C:5:9: warning: declaration of ‘i’ shadows a member of ‘S<long int>’
[-Wshadow]
    5 |   S(int i) { (void) i; }
      |     ~~~~^
shad2.C:4:7: note: shadowed declaration is here
    4 |   int i;
      |       ^
shad2.C:5:9: warning: declaration of ‘i’ shadows a member of ‘S<long int>’
[-Wshadow]
    5 |   S(int i) { (void) i; }
      |     ~~~~^
shad2.C:4:7: note: shadowed declaration is here
    4 |   int i;
      |       ^
shad2.C:5:9: warning: declaration of ‘i’ shadows a member of ‘S<long int>’
[-Wshadow]
    5 |   S(int i) { (void) i; }
      |     ~~~~^
shad2.C:4:7: note: shadowed declaration is here
    4 |   int i;
      |       ^

[Bug c++/104379] [9/10/11/12 Regression] -Wshadow warning given three times

[rguenther at suse dot de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379

--- Comment #5 from rguenther at suse dot de <rguenther at suse dot de> ---
On Fri, 4 Feb 2022, redi at gcc dot gnu.org wrote:

> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379
> 
> --- Comment #4 from Jonathan Wakely <redi at gcc dot gnu.org> ---
> So you can imagine what happens if you combine constructor clones with
> templates! :-D
> 
> template<typename T>
> struct S
> {
>   int i;
>   S(int i) { (void) i; }
> };
> 
> S<int> i(1);
> S<long> j(1);
> 
> wheeeeeeeee!

It's just really insisting on you to fix it! ;)  Somehow it feels
the shadow stuff runs at the wrong place - it should run at _parsing_,
not at instantiation [of ctor/dtor clones].

[Bug c++/104379] [9/10/11/12 Regression] -Wshadow warning given three times

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379

--- Comment #6 from Richard Biener <rguenth at gcc dot gnu.org> ---
Oh, btw - we'd also warn N times for an uninitialized variable use for example
unless the location-based diagnostic suppression gets this right now - tree or
GIMPLE no-warning flags definitely don't.

[Bug c++/104379] [9/10/11/12 Regression] -Wshadow warning given three times

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 52365
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=52365&action=edit
gcc12-pr104379.patch

Untested fix (just for -Wshadow).

[Bug c++/104379] [9/10/11/12 Regression] -Wshadow warning given three times

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379

--- Comment #8 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Jakub Jelinek <jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:71e6353e1b03021bc8bbcf4bd67a5f14d09b5fb1

commit r12-7113-g71e6353e1b03021bc8bbcf4bd67a5f14d09b5fb1
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Tue Feb 8 20:17:55 2022 +0100

    c++: Don't emit repeated -Wshadow warnings for templates/ctors [PR104379]

    The following patch suppresses extraneous -Wshadow warnings.
    On the testcase without the patch we emit 14 -Wshadow warnings,
    with the patch just 4.  It is enough to warn once e.g. during parsing of
the
    template or the abstract ctor, while previously we'd warn also on the
clones
    of the ctors and on instantiation.
    In GCC 8 and earlier we didn't warn because check_local_shadow did
      /* Inline decls shadow nothing.  */
      if (DECL_FROM_INLINE (decl))
        return;

    2022-02-08  Jakub Jelinek  <jakub@redhat.com>

            PR c++/104379
            * name-lookup.cc (check_local_shadow): When diagnosing shadowing
            of a member or global declaration, add warning suppression for
            the decl and don't warn again on it.

            * g++.dg/warn/Wshadow-18.C: New test.

[Bug c++/104379] [9/10/11 Regression] -Wshadow warning given three times

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[9/10/11/12 Regression]     |[9/10/11 Regression]
                   |-Wshadow warning given      |-Wshadow warning given
                   |three times                 |three times

--- Comment #9 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed on the trunk so far.

Note, GCC 11 and older doesn't have the warning-control.c* infrastructure, so
all we could do there is TREE_NO_WARNING.

[Bug c++/104379] [9/10/11 Regression] -Wshadow warning given three times

[rguenther at suse dot de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379

--- Comment #10 from rguenther at suse dot de <rguenther at suse dot de> ---
On Tue, 8 Feb 2022, jakub at gcc dot gnu.org wrote:

> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379
> 
> Jakub Jelinek <jakub at gcc dot gnu.org> changed:
> 
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>             Summary|[9/10/11/12 Regression]     |[9/10/11 Regression]
>                    |-Wshadow warning given      |-Wshadow warning given
>                    |three times                 |three times
> 
> --- Comment #9 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
> Fixed on the trunk so far.
> 
> Note, GCC 11 and older doesn't have the warning-control.c* infrastructure, so
> all we could do there is TREE_NO_WARNING.

Or warn from a place that doesn't visit this IL N times?

[Bug c++/104379] [10/11 Regression] -Wshadow warning given three times

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|9.5                         |10.4

--- Comment #11 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 9 branch is being closed

[Bug c++/104379] [10/11 Regression] -Wshadow warning given three times

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|10.4                        |10.5

--- Comment #12 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
GCC 10.4 is being released, retargeting bugs to GCC 10.5.

[Bug c++/104379] [11 Regression] -Wshadow warning given three times

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104379

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|10.5                        |11.5

--- Comment #13 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 10 branch is being closed.