cool I'll give it a go and let you know.

Ash


On 12 October 2016 at 17:50, Larry Gritz <l...@larrygritz.com> wrote:

> No, don't trouble yourself. Instead, I would appreciate if you could try
> applying this patch (to the current RB-1.7 head):
>
> diff --git a/CMakeLists.txt b/CMakeLists.txt
> index 00c51cf..5c22704 100644
> --- a/CMakeLists.txt
> +++ b/CMakeLists.txt
> @@ -147,6 +147,10 @@ if (CMAKE_COMPILER_IS_CLANG OR
> CMAKE_COMPILER_IS_APPLECLANG)
>          APPLECLANG_VERSION_STRING VERSION_GREATER 6.1)
>          add_definitions ("-Wno-unused-local-typedefs")
>      endif ()
> +    if (CLANG_VERSION_STRING VERSION_EQUAL 3.9 OR CLANG_VERSION_STRING
> VERSION_GREATER 3.9)
> +        # Don't warn about using unknown preprocessor symbols in #if'set
> +        add_definitions ("-Wno-expansion-to-defined")
> +    endif ()
>  endif ()
>
>
>
> I don't have 3.9 installed, so I can't easily test it myself.
>
> This should at least disable the warnings, and then we can fix the code at
> our leisure, ideally waiting until after homebrew adds a llvm39 package so
> it's easy for me to test as I go.
>
> If this works for you, I'll merge it.
>
>
> On Oct 12, 2016, at 9:37 AM, Ashley Retallack <ashley.retall...@gmail.com>
> wrote:
>
> Cool,
>
> Well I can start looking in to it if you like in  the mean time I'll use a
> lower version.
>
> Cheers
>
> Ash
>
> On 12 October 2016 at 16:15, Larry Gritz <l...@larrygritz.com> wrote:
>
>> I haven't tried clang 3.9 (though I use 3.8 regularly).
>>
>> Ugh, that's unfortunate. There are many places where we say things like
>>
>> #if A_SOMETIMES_DEFINED > 3
>>
>> instead of the more verbose
>>
>> #if defined(A_SOMETIMES_DEFINED) && A_SOMETIMES_DEFINED > 3
>>
>> I used to think the former was illegal but recently discovered that all
>> the compilers seem to accept it without complaint and figured maybe it was
>> fine all along and it was my understanding of C++ that was wrong. So I got
>> in the habit of preferring this compact form and using it frequently. But
>> this error implies that it was wrong all along and the compilers were very
>> forgiving, but perhaps in the ever-tightening warning landscape it has
>> finally fallen.
>>
>> We'll need to scour the whole codebase for constructs like that and
>> replace them with the wordier version.
>>
>> Thanks for the alert.
>>
>> > On Oct 12, 2016, at 7:27 AM, Ashley Retallack <
>> ashley.retall...@gmail.com> wrote:
>> >
>> > Hi all,
>> >
>> > Was wondering if anyone has tried building oiio 1.7 with clang 3.9.0.
>> >
>> > I've been trying and get the following:
>> >
>> > src/include/OpenImageIO/hash.h:54:37: error: macro expansion producing
>> 'defined' has undefined behavior
>> >       [-Werror,-Wexpansion-to-defined]
>> > #if OIIO_CPLUSPLUS_VERSION >= 11 || OIIO_MSVS_AT_LEAST_2013
>> >
>> > I can successfully build with gcc 4.8.4, just wanted to try and get
>> clang working.
>> >
>> > Any help much appreciated.
>> >
>> > Ash
>> > _______________________________________________
>> > Oiio-dev mailing list
>> > Oiio-dev@lists.openimageio.org
>> > http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
>>
>> --
>> Larry Gritz
>> l...@larrygritz.com
>>
>>
>> _______________________________________________
>> Oiio-dev mailing list
>> Oiio-dev@lists.openimageio.org
>> http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
>>
>
> _______________________________________________
> Oiio-dev mailing list
> Oiio-dev@lists.openimageio.org
> http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
>
>
> --
> Larry Gritz
> l...@larrygritz.com
>
>
>
> _______________________________________________
> Oiio-dev mailing list
> Oiio-dev@lists.openimageio.org
> http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
>
>
_______________________________________________
Oiio-dev mailing list
Oiio-dev@lists.openimageio.org
http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org

Reply via email to