(Potential bug) Useless warning that shouldn't exist?

Alec Teal a.teal@warwick.ac.uk
Fri Jul 12 06:17:00 GMT 2013


Thanks for the quick reply.

Turns out there was a macro, the compiler saw:
static const float (__builtin_inff());

The library must have compiled because INFINITY isn't defined, curious 
that the use does, but my IDE didn't spot that INFINITY was a macro, it 
doesn't surprise me that somewhere in the vast wxWidgets headers it's 
there. This isn't a complaint; fear not.

Sorry to waste time. Thanks Jon.

Alec

On 11/07/13 12:36, Jonathan Wakely wrote:
> On 11 July 2013 12:33, Alec Teal <a.teal@warwick.ac.uk> wrote:
>> Hey guys,
>>
>> I'm really not sure what's wrong with this, I think it may be a bug but I am
>> not sure. Hence posting it here first. These are from a shared library, that
>> compiles fine with -Wall -Wextra, the header file is available to things
>> that use the library. Particularly this has nothing to do with a  function,
>> so why is it talking of a return type of a function?
See, I was also confused about the function thing, I'd have to be pretty 
sure before firing off an email.
>>
>> Header file:
>> namespace sr {
>>
>> class Framebuffer;
>>
>> class Depthbuffer {
>> public:
>> *static const float INFINITY;*<--this line is what the warning is referring
>> to.
>> ...
>>
>> Implementation file:
>> namespace sr {
>>
>> const float Depthbuffer::INFINITY = 1.0f;
>> ....
>>
>> However when compiling a program that uses the library:
>>
>> if ! g++ -Isrc -MM src/Renderer.cpp >> build/Renderer.o.d ; then rm
>> build/Renderer.o.d ; exit 1 ; fi
>> g++ -Wall -Wextra -g -O0 -std=c++0x -gdwarf-2 -Wno-write-strings -I../libsr
>> -I/usr/lib/x86_64-linux-gnu/wx/include/gtk2-unicode-release-2.8
>> -I/usr/include/wx-2.8 -D_FILE_OFFSET_BITS=64 -D_LARGE_FILES -D__WXGTK__
>> -pthread -Isrc -c src/Renderer.cpp -o build/Renderer.o
>> In file included from ./libsr/libsr.h:12:0,
>>                   from src/Renderer.h:11,
>>                   from src/Renderer.cpp:8:
>> ./libsr/Framebuffer.h:17:21: warning: type qualifiers ignored on function
>> return type [-Wignored-qualifiers]
>>
>> The line it doesn't like is the one in bold, static const float INFINITY;
>>
>> What is GCC warning me about? What's wrong?
>> GCC 4.7.3
> That line is not a function, so there shouldn't be a warning about
> "function return type" -- are you sure that's the line it's referring
> too?  Is INFINITY a macro defined in some header so that you're not
> ompiling what you think you are?
EXACTLY, that isn't a function.
>
> Please provide a complete reproducable example, not snippets of code
> with "..." that noone else can test.
>
Overkill as a first step surely? I may have missed something trivial, I 
did miss something trivial.



More information about the Gcc-help mailing list