[Bug c++/39681] Compile error is not descriptive

[Bug c++/39681] Compile error is not descriptive

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39681

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

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

--- Comment #2 from Paolo Carlini <paolo.carlini at oracle dot com> 2011-10-02 10:07:33 UTC ---
Manuel, can I have your opinion about this one?

[Bug c++/39681] Compile error is not descriptive

[manu at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39681

--- Comment #3 from Manuel López-Ibáñez <manu at gcc dot gnu.org> 2011-10-03 10:07:49 UTC ---
(In reply to comment #2)
> Manuel, can I have your opinion about this one?

Since you ask, my opinion is that first there should be only 1 error and not
two, and bonus points if the error is something like "'foo' is not a type".
Clang says:

/tmp/webcompile/_9832_0.cc:4:18: error: expected a type
    int* p = new foo;
                 ^
1 error generated.

On the other hand, I understand that this may be difficult to fix with g++
tentative parser. The fix could be something like, once "new" is seen, then
commit to parse a new-expression. But I haven't looked at this code
specifically.

Interestingly, for:

int main()
{
    int* p = delete foo;
}

g++ says:

test.cc:3:21: error: ‘foo’ was not declared in this scope

[Bug c++/39681] Compile error is not descriptive

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39681

--- Comment #4 from Paolo Carlini <paolo.carlini at oracle dot com> 2011-10-03 11:08:39 UTC ---
Ok, thanks. Frankly I hadn't noticed the *second* error. The first one seemed
good enough to me, and quite similar to what I saw elsewhere modulo type
instead of type-specifier. So do you think it would be possible to get rid of
the stupid second error message within the current infrastructure?

[Bug c++/39681] Compile error is not descriptive

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39681

--- Comment #5 from Paolo Carlini <paolo.carlini at oracle dot com> 2011-10-03 11:13:52 UTC ---
Like, sorry about my naivete, by adding a cp_parser_skip_to_end_of_statement or
something right after the error message?!?

[Bug c++/39681] Compile error is not descriptive

[manu at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39681

Manuel López-Ibáñez <manu at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |diagnostic
             Target|i486-linux-gnu              |
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2012-05-14
               Host|i486-linux-gnu              |
     Ever Confirmed|0                           |1
              Build|i486-linux-gnu              |

--- Comment #6 from Manuel López-Ibáñez <manu at gcc dot gnu.org> 2012-05-14 11:22:43 UTC ---
This is confirmed and the error should be exactly:

error: ‘foo’ was not declared in this scope

like it happens with "delete".

Not working on it.

[Bug c++/39681] Compile error is not descriptive

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39681

--- Comment #7 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-05-14 16:07:32 UTC ---
I agree it should be better, but the analogy isn't great: "new foo" requires
foo to be a type, "delete foo" requires foo to be a variable.

[Bug c++/39681] Compile error is not descriptive

[manu at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39681

--- Comment #8 from Manuel López-Ibáñez <manu at gcc dot gnu.org> 2012-05-14 16:42:44 UTC ---
(In reply to comment #5)
> Like, sorry about my naivete, by adding a cp_parser_skip_to_end_of_statement or
> something right after the error message?!?

That may work and the result should be kind of ok. Anyway, g++ cannot do much
else with 'foo' (no spell-corrector, no searching in enclosing namespaces,
etc.).

[Bug c++/39681] Compile error is not descriptive

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39681

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|---                         |4.8.0

--- Comment #10 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-05-17 18:58:22 UTC ---
Done.

[Bug c++/39681] Compile error is not descriptive

[paolo at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39681

--- Comment #9 from paolo at gcc dot gnu.org <paolo at gcc dot gnu.org> 2012-05-17 18:54:45 UTC ---
Author: paolo
Date: Thu May 17 18:54:37 2012
New Revision: 187634

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=187634
Log:
/cp
2012-05-17  Paolo Carlini  <paolo.carlini@oracle.com>

    PR c++/39681
    * parser.c (cp_parser_new_type_id): Early return error_mark_node
    if the cp_parser_type_specifier_seq call has type_specifier_seq.type
    error_mark_node; tidy.
    (cp_parser_new_expression): Always initialize nelts to NULL_TREE to
    avoid uninitialized warnings.
    (cp_parser_init_declarator, cp_parser_late_parse_one_default_arg):
    Call cp_parser_skip_to_end_of_statement if cp_parser_initializer
    returns error_mark_node.

/testsuite
2012-05-17  Paolo Carlini  <paolo.carlini@oracle.com>

    PR c++/39681
    * g++.dg/parse/error48.C: New.
    * g++.dg/cpp0x/error8.C: Likewise.
    * g++.dg/ext/utf-cxx98.C: Adjust dg-error directive.
    * g++.dg/ext/utf-dflt2.C: Likewise.
    * g++.dg/ext/utf-gnuxx98.C: Likewise.
    * g++.dg/ext/utf-dflt.C: Likewise.
    * c-c++-common/raw-string-3.c: Likewise.
    * c-c++-common/raw-string-4.c: Likewise.

Added:
    trunk/gcc/testsuite/g++.dg/cpp0x/error8.C
    trunk/gcc/testsuite/g++.dg/parse/error49.C
Modified:
    trunk/gcc/cp/ChangeLog
    trunk/gcc/cp/parser.c
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/c-c++-common/raw-string-3.c
    trunk/gcc/testsuite/c-c++-common/raw-string-4.c
    trunk/gcc/testsuite/g++.dg/ext/utf-cxx98.C
    trunk/gcc/testsuite/g++.dg/ext/utf-dflt.C
    trunk/gcc/testsuite/g++.dg/ext/utf-dflt2.C
    trunk/gcc/testsuite/g++.dg/ext/utf-gnuxx98.C

[Bug c++/39681] Compile error is not descriptive

[paolo dot carlini at oracle dot com]

------- Comment #1 from paolo dot carlini at oracle dot com  2009-04-07 15:37 -------
FWIW, the EDG C++ front-end issues a very similar error message ;)

39681.C(3): error: expected a type specifier
      int* p = new foo;
                   ^


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39681