Mplayer Variant Help

[EmmGunn]

Thanks for the detailed response.  I don't know enough about compiling  
to come up with a universal build, so I guess I'll stick with an intel  
only build.  If I may, I have a follow up on the platform  
constraints.  So these constraints have nothing to do with what  
systems the compiled binaries will run on.  Is that right?  I seem to  
recall reading somewhere that macport builds will run on 10.4 and  
higher if I install with 10.5.  Looking at the macports guide, it  
looks like if there is no "configure.compiler" line in the port file  
and if I'm using 10.5.6, then the default is gcc-4.0 which will result  
in 10.4 and 10.5 friendly binaries.  Am understanding that correctly?   
Thanks.



On May 13, 2009, at 7:38 PM, Ryan Schmidt wrote:

> On May 13, 2009, at 12:25, EmmGunn wrote:
>
>> I'm looking at the different variants for mplayer and most of them  
>> are self explanatory, but I have a couple of questions.  I noticed  
>> this in the port file:
>>
>> 83	# configure is not autoconf
>> 84	universal_variant no
>>
>> which I took to mean that a universal variant is not available,
>
> That's correct. Because mplayer does not use a standard autoconf  
> configure script, the standard universal variant does not work with  
> it. If someone wanted to invest some time in figuring out how to  
> make mplayer compile universal, it could probably be done by writing  
> a custom universal variant for that port.
>
>> but then I see a binary_codecs variant which has something about  
>> powerpc and i386.  What does this do exactly?  Is this something  
>> equivalent to a universal variant?  I was hoping to compile a  
>> universal build of mplayer.
>
> Sorry, the port says "universal_variant no", which means someone  
> tried to build a universal binary and found that it failed, and  
> added this to the portfile to save you the trouble. If you really  
> want a universal binary, you will have to write a custom universal  
> variant for this port. If you do, please contribute it back to us so  
> that we can put it in the portfile.
>
>> Also, just out of curiosity, what are the darwin and darwin_8  
>> variants.  I couldn't find anything about them either.  Thanks in  
>> advance for any help.
>
> These are platform variants. If they are present in a portfile,  
> MacPorts automatically selects them on the appropriate platform.  
> "darwin" variants are selected on operating systems based on Darwin,  
> including Mac OS X. darwin_8 is selected on Darwin 8.x operating  
> systems, such as Mac OS X 10.4.x. Darwin 9.x would be Mac OS X  
> 10.5.x, and so on. darwin_i386 is selected if you're running Darwin  
> (or Mac OS X) on an Intel processor. darwin_powerpc is selected if  
> you're running Darwin (or Mac OS X) on a PowerPC processor. You  
> never select these variants manually; MacPorts does it for you if  
> necessary. Ports use these variants if they need different  
> instructions to build correctly in different kinds of environments.
>
> The description of the binary_codecs variant says "Enable platform- 
> specific binary codecs". Most ports in MacPorts are compiled from  
> source but sometimes bits are only available as pre-compiled  
> binaries, as is apparently the case for the codecs this variant  
> wants to enable. This variant seems to be designed to download one  
> file for PowerPC computers and a different file for Intel computers,  
> and then install their contents. However, I don't believe the  
> variant functions as intended. It checks whether variants  
> darwin_i386 or darwin_powerpc are set. Since the port does not  
> declare any variants by those names, those variants are never set,  
> hence the binary_codecs variant never does anything. This is a bug  
> in the portfile. It looks like when the binary_codecs variant was  
> set up in r20652, darwin_i386 and darwin_powerpc variants did exist  
> in the port, but they were subsequently removed without considering  
> the effect that would have on the binary_codecs variant. I fixed  
> this bug just now, but now the variant produces an error. Since I'm  
> not an MPlayer expert I filed a ticket for this problem:
>
> http://trac.macports.org/ticket/19619
>