> Oliver Stöneberg wrote:
>
> > This is simply fixed by putting those unused constants into the
> > proper defines. I attached a patch against git 70b078c.
>
> Unfortunately it breaks x86_64 build (where FLAC__CPU_X86_64 is defined, and
> FLAC__CPU_IA32 isn't). Maybe it's simpler to turn them in
Oliver Stöneberg wrote:
>> Unfortunately it breaks x86_64 build (where FLAC__CPU_X86_64 is defined, and
>> FLAC__CPU_IA32 isn't). Maybe it's simpler to turn them into #defines such as
>>
>> #define FLAC__CPUINFO_IA32_CPUID_CMOV 0x8000U
>> #define FLAC__CPUINFO_IA32_CPUID_EXTENDED_AMD
> Oliver Stöneberg wrote:
>
> >> Unfortunately it breaks x86_64 build (where FLAC__CPU_X86_64 is defined,
> >> and
> >> FLAC__CPU_IA32 isn't). Maybe it's simpler to turn them into #defines such
> >> as
> >>
> >> #define FLAC__CPUINFO_IA32_CPUID_CMOV 0x8000U
> >> #define FLAC__CPUI
Copying my report from sf.net:
The source uses a mixture of "#ifdef HAVE_CONFIG_H" and "#if HAVE_CONFIG_H"
checks. Judging from
DEFS=-DHAVE_CONFIG_H
in the configure I supposed it should be "#ifdef" in all cases. These incorrect
checks cause build errors since it tries to include the config.h
Oliver Stöneberg wrote:
> Copying my report from sf.net:
>
> The source uses a mixture of "#ifdef HAVE_CONFIG_H" and "#if HAVE_CONFIG_H"
> checks. Judging from
>
> DEFS=-DHAVE_CONFIG_H
>
> in the configure I supposed it should be "#ifdef" in all cases. These
> incorrect
> checks cause build er
> Oliver Stöneberg wrote:
>
> > Copying my report from sf.net:
> >
> > The source uses a mixture of "#ifdef HAVE_CONFIG_H" and "#if HAVE_CONFIG_H"
> > checks. Judging from
> >
> > DEFS=-DHAVE_CONFIG_H
> >
> > in the configure I supposed it should be "#ifdef" in all cases. These
> > incorrect
