[Bug c++/55671] New: -Wconversion fails to warn for sign-conversion

[Bug c++/55671] New: -Wconversion fails to warn for sign-conversion

[loose at astron dot nl]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55671

             Bug #: 55671
           Summary: -Wconversion fails to warn for sign-conversion
    Classification: Unclassified
           Product: gcc
           Version: 4.7.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: loose@astron.nl


g++ doesn't warn about the following statement when using -Wconversion, whereas
gcc does.

$ cat Wconversion-bug.c 
unsigned i = -1;

$ g++ -Wconversion -c Wconversion-bug.c
[no output]

$ gcc -Wconversion -c Wconversion-bug.c 
Wconversion-bug.c:1:1: warning: negative integer implicitly converted to
unsigned type [-Wsign-conversion]

$ g++ --version
g++ (Ubuntu/Linaro 4.7.2-2ubuntu1) 4.7.2
Copyright (C) 2012 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

[Bug c++/55671] -Wconversion fails to warn for sign-conversion

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55671

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

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

--- Comment #1 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-12-13 10:33:48 UTC ---
Looks like, if we want this is trivial to change:

Wsign-conversion
C ObjC C++ ObjC++ Var(warn_sign_conversion) LangEnabledBy(C ObjC,Wconversion)
Warn for implicit type conversions between signed and unsigned integers

Manuel, can you look into this?

[Bug c++/55671] -Wconversion fails to warn for sign-conversion

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55671

--- Comment #2 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-12-13 10:35:58 UTC ---
Surely this is by design, as documented:

  Warnings about conversions between signed and unsigned integers are
  disabled by default in C++ unless -Wsign-conversion is explicitly enabled.

[Bug c++/55671] -Wconversion fails to warn for sign-conversion

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55671

--- Comment #3 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-12-13 10:37:34 UTC ---
This is a common idiom in C++ and shouldn't warn e.g.

 static const size_type npos = -1;

[Bug c++/55671] -Wconversion fails to warn for sign-conversion

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55671

--- Comment #4 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-12-13 10:41:49 UTC ---
I agree, I suspected it was by the design. Since Manuel worked on this, let's
wait a bit before closing the Bug.

[Bug c++/55671] -Wconversion fails to warn for sign-conversion

[loose at astron dot nl]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55671

--- Comment #5 from loose at astron dot nl 2012-12-13 11:09:43 UTC ---
If this is by design, then the documentation is buggy. Here's a literal quote
form the man page of g++

       -Wconversion
           Warn for implicit conversions that may alter a value. This includes
           conversions between real and integer, like "abs (x)" when "x" is
           "double"; conversions between signed and unsigned, like "unsigned
           ui = -1"; and conversions to smaller types, like "sqrtf (M_PI)". Do
           not warn for explicit casts like "abs ((int) x)" and "ui =
           (unsigned) -1", or if the value is not changed by the conversion
           like in "abs (2.0)".  Warnings about conversions between signed and
           unsigned integers can be disabled by using -Wno-sign-conversion.

           For C++, also warn for confusing overload resolution for user-
           defined conversions; and conversions that will never use a type
           conversion operator: conversions to "void", the same type, a base
           class or a reference to them. Warnings about conversions between
           signed and unsigned integers are disabled by default in C++ unless
           -Wsign-conversion is explicitly enabled.

It explicitly gives the example "unsigned ui = -1" as an example.

Furthermore, older g++ compilers used to give this warning. I tried it with g++
4.1.2 and it properly gives a warning.

I also find it a bit counter-intuitive that gcc *does* warn about it, whereas
g++ does *not*.

[Bug c++/55671] -Wconversion fails to warn for sign-conversion

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55671

--- Comment #6 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-12-13 11:25:37 UTC ---
It gives it as an example then says "Warnings about conversions between signed
and unsigned integers are disabled by default in C++ unless -Wsign-conversion
is explicitly enabled."

You only have to read the next paragraph.

[Bug c++/55671] -Wconversion fails to warn for sign-conversion

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55671

--- Comment #7 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-12-13 11:29:24 UTC ---
This is definitely by design, as the snippet I posted also clearly shows. The
documentation could be maybe improved I agree, but it's a very general issue
that it often gives more space to C than C++ because the latter is more recent
in the development history of GCC. Maybe in this case we could repeat the
example at the end and emphasize that we do *not* warn in C++ because it's
indeed a common idiom in C++. Hey, C and C++ are two different languages,
remember, and many details in the Standards are different also in basic (eg, no
classes, no templates) areas.

[Bug c++/55671] -Wconversion fails to warn for sign-conversion

[loose at astron dot nl]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55671

--- Comment #8 from loose at astron dot nl 2012-12-13 12:03:02 UTC ---
OK, fine with me. It's only that I was caught by surprise that g++ did no
longer warn about sign conversion. I can definitely live with an explicit
-Wsign-conversion, though.

[Bug c++/55671] -Wconversion fails to warn for sign-conversion

[manu at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55671

--- Comment #9 from Manuel López-Ibáñez <manu at gcc dot gnu.org> 2012-12-13 13:24:51 UTC ---
Wsign-conversion was disabled for g++ by explicit request. I don't remember the
details and Google seems unable to find the relevant discussion in gcc-patches.

[Bug c++/55671] -Wconversion fails to warn for sign-conversion

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55671

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

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

--- Comment #10 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-12-13 23:29:42 UTC ---
Guess we can close this.