[Bug libstdc++/101937] New: error: variable-sized object may not be initialized

[Bug libstdc++/101937] New: error: variable-sized object may not be initialized

[unlvsur at live dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101937

            Bug ID: 101937
           Summary: error: variable-sized object may not be initialized
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libstdc++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: unlvsur at live dot com
  Target Milestone: ---

const char spacing[ctx._M_indent + 1] = "    ";

VLA does not initialize variable. GCC does not detect this. However when I use
clang --target=wasm32-wasi, clang complains it.

https://github.com/gcc-mirror/gcc/blob/e660441f94f2ca2ca1bd63c922c6a43737c2259e/libstdc%2B%2B-v3/src/c%2B%2B11/debug.cc#L628

../../../../gcc/libstdc++-v3/src/c++11/debug.cc:628:17: error: variable-sized
object may not be initialized
            const char spacing[ctx._M_indent + 1] = "    ";
                       ^                            ~~~~~~

[Bug libstdc++/101937] error: variable-sized object may not be initialized

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101937

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
but it's 'const' so it must be initialized?

[Bug libstdc++/101937] error: variable-sized object may not be initialized

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101937

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

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

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
VLAs are an extension in C++, so there is no standard that says whether they
can or can't be initialized.  C99 says they can't, but that doesn't mean they
can't be in C++...

[Bug libstdc++/101937] error: variable-sized object may not be initialized

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101937

--- Comment #3 from Jonathan Wakely <redi at gcc dot gnu.org> ---
It's not a VLA though, _M_indent is an enumerator with the value 4, and
ctx._M_index + 1 is 5.

[Bug libstdc++/101937] error: variable-sized object may not be initialized

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101937

--- Comment #4 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Clang accepts this:

struct S {
  enum { _M_indent = 4 };
};

int f(S s)
{
  const char str[s._M_indent + 1] = "    ";
  return sizeof(str);
}

But rejects the same program with a reference parameter:

struct S {
  enum { _M_indent = 4 };
};

int f(S& s)
{
  const char str[s._M_indent + 1] = "    ";
  return sizeof(str);
}


This will fix it for clang:

--- a/libstdc++-v3/src/c++11/debug.cc
+++ b/libstdc++-v3/src/c++11/debug.cc
@@ -625,8 +625,8 @@ namespace
        // If this isn't the first line, indent
        if (ctx._M_column == 1 && !ctx._M_first_line)
          {
-           const char spacing[ctx._M_indent + 1] = "    ";
-           print_raw(ctx, spacing, ctx._M_indent);
+           const char spacing[PrintContext::_M_indent + 1] = "    ";
+           print_raw(ctx, spacing, PrintContext::_M_indent);
          }

        int written = fprintf(stderr, "%.*s", (int)length, word);

[Bug libstdc++/101937] error: variable-sized object may not be initialized

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101937

--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Do we need to work around clang bugs here though?  In headers I can understand
it, but this is in the libstdc++.{so,a} only code.

[Bug libstdc++/101937] error: variable-sized object may not be initialized

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101937

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2021-08-16
           Assignee|unassigned at gcc dot gnu.org      |redi at gcc dot gnu.org
             Status|UNCONFIRMED                 |ASSIGNED
     Ever confirmed|0                           |1

--- Comment #6 from Jonathan Wakely <redi at gcc dot gnu.org> ---
The problem is that ctx._M_indent is a member access expression through a
reference, which is not allowed in constant expressions. This is a silly
restriction that https://wg21.link/p2280r2 aims to change.

[Bug libstdc++/101937] error: variable-sized object may not be initialized

[unlvsur at live dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101937

--- Comment #7 from cqwrteur <unlvsur at live dot com> ---
(In reply to Jakub Jelinek from comment #5)
> Do we need to work around clang bugs here though?  In headers I can
> understand it, but this is in the libstdc++.{so,a} only code.

well i love gcc, however, unfortunately, gcc does not provide webassembly
backend ( wasm32/64-wasi).

The only thing i could do is to use clang to compile libstdc++ and get partial
experience with it.

hope you guys can work on providing backend in the future.

[Bug libstdc++/101937] error: variable-sized object may not be initialized

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101937

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

https://gcc.gnu.org/g:6c25932ac399423b09b730fb8f894ada568deb2a

commit r12-2933-g6c25932ac399423b09b730fb8f894ada568deb2a
Author: Jonathan Wakely <jwakely@redhat.com>
Date:   Mon Aug 16 15:35:58 2021 +0100

    libstdc++: Use qualified-id for class member constant [PR101937]

    The expression ctx._M_indent is not a constant expression when ctx is a
    reference parameter, even though _M_indent is an enumerator. Rename it
    to _S_indent to be consistent with our conventions, and refer to it as
    PrintContext::_S_indent to be valid C++ code (at least until P2280 is
    accepted as a DR).

    Signed-off-by: Jonathan Wakely <jwakely@redhat.com>

    libstdc++-v3/ChangeLog:

            PR libstdc++/101937
            * src/c++11/debug.cc (PrintContext::_M_indent): Replace with a
            static data member.
            (print_word): Use qualified-id to access it.

[Bug libstdc++/101937] error: variable-sized object may not be initialized

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101937

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |12.0
         Resolution|---                         |FIXED
             Status|ASSIGNED                    |RESOLVED

--- Comment #9 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Fixed on trunk.

[Bug libstdc++/101937] error: variable-sized object may not be initialized

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101937

--- Comment #10 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Jonathan Wakely from comment #6)
> The problem is that ctx._M_indent is a member access expression through a
> reference, which is not allowed in constant expressions. This is a silly
> restriction that https://wg21.link/p2280r2 aims to change.

Just FYI the GCC bug accepting this code is PR 72756.