[macports-ports] branch master updated: ffmpeg: force gcc6 on PPC

[Ken Cunningham]

The fallbacks on PPC don't work correctly, I find.

If I were to blacklist gcc42, MacPorts would lead in almost all cases to building it with clang-3.4, which always fails. 

Also, there's no point falling back to any gcc prior to gcc6.

So what I saw as options was to blacklist:

all the gcc versions individually prior to gcc6
all the {clang} versions
all the macports-clang-* versions

and then  add fallbacks to macports-gcc-6


-- or -- 

just clear the blacklisting and whitelist macports-gcc-6 like we do in the cxx11 1.1 PortGroup

It just seemed much more straightforward to do the latter.

I'm always open to better plans and opinions, though.

Ken




On 2017-10-20, at 7:08 PM, Ryan Schmidt wrote:

> 
> On Oct 20, 2017, at 20:49, Ken wrote:
> 
>> Ken (kencu) pushed a commit to branch master
>> in repository macports-ports.
>> 
>> 
>> https://github.com/macports/macports-ports/commit/e1f1b6b20678fed08e81f8a65ea0062f59a1f64c
>> 
>> The following commit(s) were added to refs/heads/master by this push:
>> 
>>     new e1f1b6b  ffmpeg: force gcc6 on PPC
>> 
>> e1f1b6b is described below
>> 
>> 
>> commit e1f1b6b20678fed08e81f8a65ea0062f59a1f64c
>> 
>> Author: Ken Cunningham <kencu at macports.org>
>> AuthorDate: Fri Oct 20 18:49:11 2017 -0700
>> 
>>    ffmpeg: force gcc6 on PPC
>> 
>>    closes: https://trac.macports.org/ticket/55124
>> 
>> ---
>> multimedia/ffmpeg/Portfile | 8 ++++++++
>> 1 file changed, 8 insertions(+)
>> 
>> diff --git a/multimedia/ffmpeg/Portfile b/multimedia/ffmpeg/Portfile
>> index bdc0919..61ec6e3 100644
>> --- a/multimedia/ffmpeg/Portfile
>> +++ b/multimedia/ffmpeg/Portfile
>> @@ -106,6 +106,14 @@ if {[lsearch [get_canonical_archs] i386] != -1} {
>>     compiler.blacklist-append {clang < 422.1.7}
>> }
>> 
>> +platform darwin powerpc {
>> +    # https://trac.macports.org/ticket/55124
>> +    # requires newer compiler than gcc4.2 to build
>> +    compiler.blacklist
>> +    compiler.whitelist    macports-gcc-6
>> +    universal_variant     no
>> +}
> 
> 
> Is there a specific reason why you did it this way, instead of what we usually do, which is to blacklist the compilers that don't work and let MacPorts choose the next-best alternative? I know gcc* isn't in the list of compilers MacPorts will fall back to. Do we maybe just need to add gcc6 to the fallback list on PowerPC?
> 
> 
>