[Bug c++/113760] New: gcc rejects valid empty-declaration in pedantic mode

[Bug c++/113760] New: gcc rejects valid empty-declaration in pedantic mode

[tanksherman27 at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

            Bug ID: 113760
           Summary: gcc rejects valid empty-declaration in pedantic mode
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: tanksherman27 at gmail dot com
  Target Milestone: ---

In https://github.com/openjdk/jdk/pull/17687#issuecomment-1926154325 we
discovered a possible g++ bug where g++ rejects empty semicolons in pedantic
mode, this is rather evident in code like

#define DEBUG_ONLY(code) code;

DEBUG_ONLY(foo());

which will fire a -Werror warning even in C++11 and above, where such empty
declarations are allowed

[Bug c++/113760] gcc rejects valid empty-declaration in pedantic mode

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2024-02-05
             Status|UNCONFIRMED                 |WAITING
     Ever confirmed|0                           |1

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I can't reproduce this at all.
I tried:
```
void foo();
#define DEBUG_ONLY(code) code;

void f() {
DEBUG_ONLY(foo());
}
```

The only time I can reproduce an error is:
```
;
```

And with `-std=c++98 -W -Wall -pedantic-errors`.

Can you provide a full example where you are getting into an issue?

[Bug c++/113760] gcc rejects valid empty-declaration in pedantic mode

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           See Also|                            |https://gcc.gnu.org/bugzill
                   |                            |a/show_bug.cgi?id=96068

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Or is it just a dup of bug 96068 ? Which was fixed in GCC 11.1.0. GCC 10.5.0
was the last released of the GCC 10 branch. And GCC 9 is also in this mode.

[Bug c++/113760] gcc rejects valid empty-declaration in pedantic mode

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
From JDK bug report:
> It looks like it only removed
> the warnings for empty declarations at namespace scope. I couldn't find
> anything for other cases, including empty class member declarations.

Yes because the C++ defect report was only for `Spurious semicolons at
namespace scope should be allowed`.  See
https://www.open-std.org/jtc1/sc22/wg21/docs/cwg_defects.html#569 .


```
struct f
{
  int t;  ;
};
```

Is not allowed by the C++ standard currently and is a GCC extension, maybe it
should have a seperate flag to control that but I am not 100% sure.

[Bug c++/113760] gcc rejects valid empty-declaration in pedantic mode

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

--- Comment #4 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Please provide a proper bug report not a macro definition out of context and a
URL.

https://gcc.gnu.org/bugs/

[Bug c++/113760] gcc rejects valid empty-declaration in pedantic mode

[kab at acm dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

kab at acm dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |kab at acm dot org

--- Comment #5 from kab at acm dot org ---
(In reply to Andrew Pinski from comment #3)
> From JDK bug report:
> > It looks like it only removed
> > the warnings for empty declarations at namespace scope. I couldn't find
> > anything for other cases, including empty class member declarations.
> 
> Yes because the C++ defect report was only for `Spurious semicolons at
> namespace scope should be allowed`.  See
> https://www.open-std.org/jtc1/sc22/wg21/docs/cwg_defects.html#569 .
> 
> 
> ```
> struct f
> {
>   int t;  ;
> };
> ```
> 
> Is not allowed by the C++ standard currently and is a GCC extension, maybe
> it should have a seperate flag to control that but I am not 100% sure.

It's C++14 that the openjdk currently cares about. C++14 added
"empty-declaration" to the syntax for "member-declaration".
C++14 9.2 [class.mem]
https://www.open-std.org/jtc1/sc22/wg21/docs/cwg_defects.html#1693

The above "struct f" declaration warns with C++14 and -Wpedantic, but
shouldn't.

Tested with that struct as the only contents of "test.cpp" and this command
line for g++13.2.0:
g++ -c -Wpedantic -std=c++14 -o test.o test.cpp

[Bug c++/113760] [DR1693] gcc rejects valid empty-declaration in pedantic mode

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|gcc rejects valid           |[DR1693] gcc rejects valid
                   |empty-declaration in        |empty-declaration in
                   |pedantic mode               |pedantic mode
           Keywords|                            |rejects-valid
             Status|WAITING                     |NEW

--- Comment #6 from Jonathan Wakely <redi at gcc dot gnu.org> ---
OK, so a proper bug report would be that there's a diagnostic with -std=c++14
-pedantic for this code:

struct f
{
  int t; ;
};

Confirmed.

[Bug c++/113760] [DR1693] gcc rejects valid empty-declaration in pedantic mode

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

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

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

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
g++ emits 4 errors on
struct S
{
  void foo () {}
  void bar () {};
  void baz () = delete;
  void qux () = delete;
  ;
  void corge () = delete;
  ;
  ;
  int s;
  ;
};
;
;
(one after qux, 2 after corge, one after s), clang++ -pedantic-errors
-std=c++23 including trunk 2 (one after corge, one after s), so neither
implements the DR.  Strangely, with -pedantic-errors -std=c++23 -Wextra-semi it
warns 7 times but doesn't error (which is I think the desirable state).
Now, g++ with -pedantic-errors -std=c++23 -Wextra-semi emits just one warning
on the ; after bar and still the 4 errors.

That said, -Wextra-semi in GCC is documented that way:
     Warn about redundant semicolons after in-class function
     definitions.
and clang doesn't bother to document it at all (at least haven't found it).

So, shall we change documentation of -Wextra-semi and say change
      /* A declaration consisting of a single semicolon is invalid
       * before C++11.  Allow it unless we're being pedantic.  */
      if (cxx_dialect < cxx11)
        pedwarn (input_location, OPT_Wpedantic, "extra %<;%>");
to else warning (OPT_Wextra_semi, "extra %<;%>"); etc.?
Then there is
          if (!in_system_header_at (token->location))
            {
              gcc_rich_location richloc (token->location);
              richloc.add_fixit_remove ();
              pedwarn (&richloc, OPT_Wpedantic, "extra %<;%>");
            }
should that be similarly if (cxx_dialect < cxx11) pedwarn; else warning ?

Or, if we want to change this already for GCC 14, do that if (cxx_dialect <
cxx11) part
just before the last pedwarn above and add the else warning for GCC 15?

[Bug c++/113760] [DR1693] gcc rejects valid empty-declaration in pedantic mode

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

--- Comment #8 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #7)
> g++ emits 4 errors on
> struct S
> {
>   void foo () {}
>   void bar () {};
>   void baz () = delete;
>   void qux () = delete;
>   ;
>   void corge () = delete;
>   ;
>   ;
>   int s;
>   ;
> };
> ;
> ;
> (one after qux, 2 after corge, one after s), clang++ -pedantic-errors
> -std=c++23 including trunk 2 (one after corge, one after s), so neither
> implements the DR.  Strangely, with -pedantic-errors -std=c++23 -Wextra-semi
> it warns 7 times but doesn't error (which is I think the desirable state).
> Now, g++ with -pedantic-errors -std=c++23 -Wextra-semi emits just one
> warning on the ; after bar and still the 4 errors.
> 
> That said, -Wextra-semi in GCC is documented that way:
>      Warn about redundant semicolons after in-class function
>      definitions.
> and clang doesn't bother to document it at all (at least haven't found it).
> 
> So, shall we change documentation of -Wextra-semi and say change
>       /* A declaration consisting of a single semicolon is invalid
>        * before C++11.  Allow it unless we're being pedantic.  */
>       if (cxx_dialect < cxx11)
>         pedwarn (input_location, OPT_Wpedantic, "extra %<;%>");
> to else warning (OPT_Wextra_semi, "extra %<;%>"); etc.?

Yeah, I'd say so.

> Then there is
>           if (!in_system_header_at (token->location))
>             {
>               gcc_rich_location richloc (token->location);
>               richloc.add_fixit_remove ();
>               pedwarn (&richloc, OPT_Wpedantic, "extra %<;%>");
>             }
> should that be similarly if (cxx_dialect < cxx11) pedwarn; else warning ?

I think so.  The < cxx11 is definitely missing here.

> Or, if we want to change this already for GCC 14, do that if (cxx_dialect <
> cxx11) part
> just before the last pedwarn above and add the else warning for GCC 15?

I would do it all at once.

I guess:

{} -> no warnings in any dialect
-pedantic -> pedwarns only in C++03
-Wextra-semi -> all warnings in all dialects
-std=c++11 -pedantic -Wextra-semi -> only -Wextra-semi warnings
-std=c++03 -pedantic -Wextra-semi -> pedwarns
-std=c++11 -pedantic -Wno-extra-semi -> no warnings
-std=c++03 -pedantic -Wno-extra-semi -> no warnings
-std=c++03 -pedantic-errors -Wextra-semi -> errors (?)
-std=c++11 -pedantic-errors -Wextra-semi -> warnings

[Bug c++/113760] [DR1693] gcc rejects valid empty-declaration in pedantic mode

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

--- Comment #9 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
And:
-pedantic-errors -> errors only in C++03

[Bug c++/113760] [DR1693] gcc rejects valid empty-declaration in pedantic mode

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |mpolacek at gcc dot gnu.org
             Status|NEW                         |ASSIGNED

[Bug c++/113760] [DR1693] gcc rejects valid empty-declaration in pedantic mode

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

--- Comment #10 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
And we should also warn in C++98 with -Wc++11-extensions for an extra ';'
outside of a function I suppose...

[Bug c++/113760] [DR1693] gcc rejects valid empty-declaration in pedantic mode

[harald at gigawatt dot nl]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

Harald van Dijk <harald at gigawatt dot nl> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |harald at gigawatt dot nl

--- Comment #11 from Harald van Dijk <harald at gigawatt dot nl> ---
(In reply to Marek Polacek from comment #8)
> -std=c++03 -pedantic-errors -Wextra-semi -> errors (?)

Speaking as a user: that makes sense to me, but I would also expect:

-std=c++03 -pedantic-errors -Wno-error=extra-semi -> warnings

[Bug c++/113760] [DR1693] gcc rejects valid empty-declaration in pedantic mode

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

--- Comment #12 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
Thank for your comment.  In the end I went with

-std=c++03 -pedantic-errors -Wextra-semi -> warnings
-std=c++03 -pedantic -Wextra-semi -> warnings (not pedwarn)

based on the principle that a more specific option overrides a more general
option.  This is also what clang++ does.  Granted, -Wvla in C doesn't behave
like that...

[Bug c++/113760] [DR1693] gcc rejects valid empty-declaration in pedantic mode

[harald at gigawatt dot nl]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

--- Comment #13 from Harald van Dijk <harald at gigawatt dot nl> ---
(In reply to Marek Polacek from comment #12)
> Thank for your comment.  In the end I went with
> 
> -std=c++03 -pedantic-errors -Wextra-semi -> warnings
> -std=c++03 -pedantic -Wextra-semi -> warnings (not pedwarn)
> 
> based on the principle that a more specific option overrides a more general
> option.  This is also what clang++ does.  Granted, -Wvla in C doesn't behave
> like that...

That also makes sense. The more specific option overriding a more general
option is also the reasoning why I expect no error with -pedantic-errors
-Wno-error=extra-semi.

[Bug c++/113760] [DR1693] gcc rejects valid empty-declaration in pedantic mode

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

--- Comment #14 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
My current patch appears to handle that correctly:

$ ./cc1plus -quiet q.C -pedantic-errors -std=c++98 -Wno-error=extra-semi
q.C:3:3: warning: extra ‘;’ outside of a function only allowed in C++11
[-Wextra-semi]
    3 | };;
      |   ^
      |   -

I'll add a test for that combination.

[Bug c++/113760] [DR1693] gcc rejects valid empty-declaration in pedantic mode

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch

--- Comment #15 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
Patch posted.  I don't foresee it being part of GCC 14.
https://gcc.gnu.org/pipermail/gcc-patches/2024-February/645494.html

[Bug c++/113760] [DR1693] gcc rejects valid empty-declaration in pedantic mode

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

--- Comment #16 from GCC Commits <cvs-commit at gcc dot gnu.org> ---
The trunk branch has been updated by Marek Polacek <mpolacek@gcc.gnu.org>:

https://gcc.gnu.org/g:6fec511f2d23cc70ab29d1ba79c2415ab51bcb60

commit r14-8967-g6fec511f2d23cc70ab29d1ba79c2415ab51bcb60
Author: Marek Polacek <polacek@redhat.com>
Date:   Tue Feb 13 16:21:32 2024 -0500

    c++: adjust the extra ; warning [PR113760]

    A minimal fix to quash an extra ; warning.  I have a more complete
    patch for GCC 15.

            DR 1693
            PR c++/113760

    gcc/cp/ChangeLog:

            * parser.cc (cp_parser_member_declaration): Only pedwarn about an
extra
            semicolon in C++98.

    gcc/testsuite/ChangeLog:

            * g++.dg/semicolon-fixits.C: Run in C++98 only.
            * g++.dg/warn/pedantic2.C: Adjust dg-warning.
            * g++.old-deja/g++.jason/parse11.C: Adjust dg-error.
            * g++.dg/DRs/dr1693-1.C: New test.
            * g++.dg/DRs/dr1693-2.C: New test.

[Bug c++/113760] [DR1693] gcc rejects valid empty-declaration in pedantic mode

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

--- Comment #17 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
Partially fixed for GCC 14.  Leaving this open for more changes in GCC 15.

[Bug c++/113760] [DR1693] gcc rejects valid empty-declaration in pedantic mode

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

--- Comment #18 from GCC Commits <cvs-commit at gcc dot gnu.org> ---
The trunk branch has been updated by Marek Polacek <mpolacek@gcc.gnu.org>:

https://gcc.gnu.org/g:0b3eac4b54a52bf206b88743d1e987badc97cff4

commit r15-522-g0b3eac4b54a52bf206b88743d1e987badc97cff4
Author: Marek Polacek <polacek@redhat.com>
Date:   Mon Feb 12 19:36:16 2024 -0500

    c++: DR 569, DR 1693: fun with semicolons [PR113760]

    Prompted by c++/113760, I started looking into a bogus "extra ;"
    warning in C++11.  It quickly turned out that if I want to fix
    this for good, the fix will not be so small.

    This patch touches on DR 569, an extra ; at namespace scope should
    be allowed since C++11:

      struct S {
      };
      ; // pedwarn in C++98

    It also touches on DR 1693, which allows superfluous semicolons in
    class definitions since C++11:

      struct S {
        int a;
        ; // pedwarn in C++98
      };

    Note that a single semicolon is valid after a member function definition:

      struct S {
        void foo () {}; // only warns with -Wextra-semi
      };

    There's a new function maybe_warn_extra_semi to handle all of the above
    in a single place.  So now they all get a fix-it hint.

    -Wextra-semi turns on all "extra ;" diagnostics.  Currently, options
    like -Wc++11-compat or -Wc++11-extensions are not considered.

            DR 1693
            PR c++/113760
            DR 569

    gcc/c-family/ChangeLog:

            * c.opt (Wextra-semi): Initialize to -1.

    gcc/cp/ChangeLog:

            * parser.cc (extra_semi_kind): New.
            (maybe_warn_extra_semi): New.
            (cp_parser_declaration): Call maybe_warn_extra_semi.
            (cp_parser_member_declaration): Likewise.

    gcc/ChangeLog:

            * doc/invoke.texi: Update -Wextra-semi documentation.

    gcc/testsuite/ChangeLog:

            * g++.dg/diagnostic/semicolon1.C: New test.
            * g++.dg/diagnostic/semicolon10.C: New test.
            * g++.dg/diagnostic/semicolon11.C: New test.
            * g++.dg/diagnostic/semicolon12.C: New test.
            * g++.dg/diagnostic/semicolon13.C: New test.
            * g++.dg/diagnostic/semicolon14.C: New test.
            * g++.dg/diagnostic/semicolon15.C: New test.
            * g++.dg/diagnostic/semicolon16.C: New test.
            * g++.dg/diagnostic/semicolon17.C: New test.
            * g++.dg/diagnostic/semicolon2.C: New test.
            * g++.dg/diagnostic/semicolon3.C: New test.
            * g++.dg/diagnostic/semicolon4.C: New test.
            * g++.dg/diagnostic/semicolon5.C: New test.
            * g++.dg/diagnostic/semicolon6.C: New test.
            * g++.dg/diagnostic/semicolon7.C: New test.
            * g++.dg/diagnostic/semicolon8.C: New test.
            * g++.dg/diagnostic/semicolon9.C: New test.

[Bug c++/113760] [DR1693] gcc rejects valid empty-declaration in pedantic mode

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113760

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |FIXED

--- Comment #19 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
Ought to be fully fixed now.