[Bug c++/49952] New: Unicode literals do not generate errors as prescribed by the FDIS standard

[Bug c++/49952] New: Unicode literals do not generate errors as prescribed by the FDIS standard

[z0sh at sogetthis dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49952

           Summary: Unicode literals do not generate errors as prescribed
                    by the FDIS standard
           Product: gcc
           Version: 4.6.1
            Status: UNCONFIRMED
          Severity: trivial
          Priority: P3
         Component: c++
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: z0sh@sogetthis.com
              Host: Linux x86


Referring to the standard:
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2007/n2442.htm

The character literal \UNNNNNNNN must only accept characters in the range
0-0x10FFFF, excluding surrogates. However, GCC allows 31-bit values above
0x10FFFF. To wit, the following compiles:

    char32_t s[] = U"\U0010FFFF\U7FFFFFFF";

It may be that the actual wording of the FDIS (2.3.2, p.19) is more relaxed
than in the reference I gave above and that this behaviour is in fact
intentional, but I thought I bring it up anyway.

[Bug c++/49952] [C++0x] Unicode literals do not generate errors as prescribed by the FDIS standard

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49952

Paolo Carlini <paolo.carlini at oracle dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |kris.van.hees at oracle dot
                   |                            |com

--- Comment #1 from Paolo Carlini <paolo.carlini at oracle dot com> 2011-08-02 22:32:10 UTC ---
Kris, are you willing to triage this PR?

[Bug c++/49952] [C++0x] Unicode literals do not generate errors as prescribed by the FDIS standard

[joseph at codesourcery dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49952

--- Comment #2 from joseph at codesourcery dot com <joseph at codesourcery dot com> 2011-08-03 11:19:01 UTC ---
C and C++ reference ISO 10646 instead of Unicode, meaning that it is 
natural and proper for the full ISO 10646 range of values to be accepted 
instead of the restricted Unicode range.  N3291 does appear to have this 
restriction on char32_t string (but not character) literals; C1X does not.

[Bug c++/49952] [C++0x] Unicode literals do not generate errors as prescribed by the FDIS standard

[z0sh at sogetthis dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49952

--- Comment #3 from Kerrek SB <z0sh at sogetthis dot com> 2011-08-03 11:36:41 UTC ---
Maybe it could trigger a warning in -pedantic mode?

[Bug c++/49952] [C++0x] Unicode literals do not generate errors as prescribed by the FDIS standard

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49952

--- Comment #4 from Paolo Carlini <paolo.carlini at oracle dot com> 2011-08-03 12:00:30 UTC ---
Adding a warning would be easy.

[Bug c++/49952] [C++0x] Unicode literals do not generate errors as prescribed by the FDIS standard

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=49952

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2021-12-02
           Keywords|                            |diagnostic
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW

--- Comment #5 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
With -std=c++20, we do warn:
<source>:2:21: warning: \U7FFFFFFF is outside the UCS codespace
    2 |     char32_t  s[] = U"\U0010FFFF\U7FFFFFFF";
      |                     ^~~~~~~~~~~~~~~~~~~~~~~

This warning was implemented in r10-3414-g0900e29cdbc5.
I wonder if we should just enable it for all C++ standards and above of C++20+.