[Bug c++/101710] New: Sometimes constexpr cannot be used as constexpr

[Bug c++/101710] New: Sometimes constexpr cannot be used as constexpr

[steve_green at qq dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101710

            Bug ID: 101710
           Summary: Sometimes constexpr cannot be used as constexpr
           Product: gcc
           Version: 11.2.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: steve_green at qq dot com
  Target Milestone: ---

Created attachment 51231
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=51231&action=edit
.ii file & .s file & output

I want to make a minimal reproduction of this ICE, but it seems I can't
So allow me to post a link to the project, sorry


You can try to compile this file
https://github.com/ELC-lang/ELC/blob/master/parts/header_file/test/forICE.cpp

In MSVC, the compilation of this file will stop due to some ICE, but it is
different from ICEs that GCC have 
The problem with GCC is that all "error: expected primary-expression before ‘X’
token" can be compiled in MSVC

[Bug c++/101710] Sometimes constexpr cannot be used as constexpr

[steve_green at qq dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101710

steve02081504 <steve_green at qq dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #51231|0                           |1
        is obsolete|                            |

--- Comment #1 from steve02081504 <steve_green at qq dot com> ---
Created attachment 51232
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=51232&action=edit
.ii file & .s file & output

[Bug c++/101710] Sometimes constexpr cannot be used as constexpr

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101710

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-on-valid-code

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Can you re-upload the zip file as it seems be corrupted?

[Bug c++/101710] Sometimes constexpr cannot be used as constexpr

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101710

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #2)
> Can you re-upload the zip file as it seems be corrupted?

Never mind, I was able to extract it, I had to use winrar rather than the
builtin unziper in windows or Info-ZIP unzip on Linux.

[Bug c++/101710] Sometimes constexpr cannot be used as constexpr

[steve_green at qq dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101710

--- Comment #4 from steve02081504 <steve_green at qq dot com> ---
Created attachment 51235
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=51235&action=edit
.ii file & .s file & output - Deflate64

Deflate64 zip added.

[Bug c++/101710] Sometimes constexpr cannot be used as constexpr

[steve_green at qq dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101710

--- Comment #5 from steve02081504 <steve_green at qq dot com> ---
(In reply to Andrew Pinski from comment #3)
> (In reply to Andrew Pinski from comment #2)
> > Can you re-upload the zip file as it seems be corrupted?
> 
> Never mind, I was able to extract it, I had to use winrar rather than the
> builtin unziper in windows or Info-ZIP unzip on Linux.

Sorry, I think I used LZMA format in the compressed package, which is probably
not supported by the standard ZIP format.

[Bug c++/101710] Sometimes constexpr cannot be used as constexpr

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101710

--- Comment #6 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Changing:
  if
constexpr(type_info<T>.has_attribute_helper<can_t_use_default_null_ptr>()&&type_info<T>.not_has_has_attribute_helper<force_use_default_null_ptr>())

to:
  if constexpr(type_info<T>.template
has_attribute_helper<can_t_use_default_null_ptr>()&&type_info<T>.template
not_has_has_attribute_helper<force_use_default_null_ptr>())

Fixes many of the errors
Note the use of template.

Note there are cases where GCC does not realize the template keyword is not
needed but I don't know if it is actually needed in this case or not.  I
suspect it is.

[Bug c++/101710] Sometimes constexpr cannot be used as constexpr

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101710

--- Comment #7 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #6)
> Changing:
>   if
> constexpr(type_info<T>.
> has_attribute_helper<can_t_use_default_null_ptr>()&&type_info<T>.
> not_has_has_attribute_helper<force_use_default_null_ptr>())
> 
> to:
>   if constexpr(type_info<T>.template
> has_attribute_helper<can_t_use_default_null_ptr>()&&type_info<T>.template
> not_has_has_attribute_helper<force_use_default_null_ptr>())
> 
> Fixes many of the errors
> Note the use of template.

This shows up in many places in your files.

There was two other errors I noticed:
* aligned_alloc is not defined; mingw does not define aligned_alloc I think.
* is_pointer<T> should have been std::is_pointer_v<T> as far as I can tell.

[Bug c++/101710] Sometimes constexpr cannot be used as constexpr

[steve_green at qq dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101710

steve02081504 <steve_green at qq dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |INVALID
             Status|UNCONFIRMED                 |RESOLVED

--- Comment #8 from steve02081504 <steve_green at qq dot com> ---
(In reply to Andrew Pinski from comment #6)
> Changing:
>   if
> constexpr(type_info<T>.
> has_attribute_helper<can_t_use_default_null_ptr>()&&type_info<T>.
> not_has_has_attribute_helper<force_use_default_null_ptr>())
> 
> to:
>   if constexpr(type_info<T>.template
> has_attribute_helper<can_t_use_default_null_ptr>()&&type_info<T>.template
> not_has_has_attribute_helper<force_use_default_null_ptr>())
> 
> Fixes many of the errors
> Note the use of template.
> 
> Note there are cases where GCC does not realize the template keyword is not
> needed but I don't know if it is actually needed in this case or not.  I
> suspect it is.

So I just need to add "template" to these errors?
Well, although it doesn't look very beautiful, I think I latched on
This is my code's problem, not GCC's ICE, right
thank you