public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "redi at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/114265] Unhelpful message when var name is also a struct name
Date: Thu, 07 Mar 2024 11:57:50 +0000	[thread overview]
Message-ID: <bug-114265-4-ajroRbKkxz@http.gcc.gnu.org/bugzilla/> (raw)
In-Reply-To: <bug-114265-4@http.gcc.gnu.org/bugzilla/>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114265

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2024-03-07
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1

--- Comment #1 from Jonathan Wakely <redi at gcc dot gnu.org> ---
But it's valid to define Struct as a variable after also defining it as a type,
the problem is only when you later refer to it as a type. That's why the error
talks about the 'fn' being redeclared. When it sees `int fn(Struct*)` the
second time, the language grammar says you are declaring an int variable, and
initializing it with an arithmetic expression involving the Struct variable
multiplied by a missing operand.

The code would be valid if the use of Struct after its redefinition used an
elaborated-type-specifier to refer to it (i.e. included the 'struct' keyword):

int fn(struct Struct *)
{ 
    return 0;
}

Maybe it would be possible for the "redeclared as a different kind of entity"
to try to parse `int fn(Struct*)` again, excluding non-types from name lookup
(except in unevaluated operands and default arguments?) and then if that
succeeds, adda fix-it hint for the elaborated-type-specifier (with either
'struct', 'class', 'union', or 'enum' as appropriate):

note: use 'struct Struct' to refer to the type
int fn(struct Struct *)
       ^~~~~~



For comparison, other compilers give very similar results:

Clang gives:

s.cc:7:5: error: redefinition of 'fn' as different kind of symbol
int fn(Struct *)
    ^
s.cc:3:5: note: previous definition is here
int fn(Struct *);
    ^
s.cc:7:16: error: expected expression
int fn(Struct *)
               ^
s.cc:7:17: error: expected ';' after top level declarator
int fn(Struct *)
                ^
                ;
3 errors generated.


EDG gives:


"s.cc", line 7: error: declaration is incompatible with "int fn(Struct *)"
          (declared at line 3)
  int fn(Struct *)
      ^

"s.cc", line 7: error: expected an expression
  int fn(Struct *)
                 ^

"s.cc", line 8: error: expected a ";"
  { 
  ^

3 errors detected in the compilation of "s.cc".



MSVC seems the least helpful:

<source>(7): error C2059: syntax error: ')'
<source>(8): error C2143: syntax error: missing ';' before '{'
<source>(8): error C2447: '{': missing function header (old-style formal list?)

      reply	other threads:[~2024-03-07 11:57 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-03-07  9:12 [Bug c++/114265] New: " denis.campredon at gmail dot com
2024-03-07 11:57 ` redi at gcc dot gnu.org [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=bug-114265-4-ajroRbKkxz@http.gcc.gnu.org/bugzilla/ \
    --to=gcc-bugzilla@gcc.gnu.org \
    --cc=gcc-bugs@gcc.gnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).