iOS cross compiling support

[Bradley Giesbrecht]

On Jan 25, 2011, at 1:58 PM, James Gregurich wrote:

> There is one more issue I want to address before I submit.
>
> To configure icu for iOS, you have to have --host=arm-apple-darwin  
> as one of the options. I'd like to automatically and generically  
> create the label to pass to --host.  However, I'm having a hard time  
> finding documentation on how these platform strings are to be  
> formatted and what the options are in them. I don't even know what  
> terminology to use to search for info on them. What are they called  
> so I can go searching?  Right now, I just have arm-apple-darwin  
> hardcoded into the port file.

How about "platform"?

man portfile

--
Brad

> On Jan 24, 2011, at 7:02 PM, Landon J Fuller wrote:
>
>>
>> On Jan 24, 2011, at 9:37 PM, James Gregurich wrote:
>>
>>> hi.
>>>
>>> I've done the proof-of-concept work on extending macports to  
>>> handle cross-compiling so that it can be used to maintain 3rd  
>>> party libs for iOS development. I have successfully built ICU  
>>> against the iPhone sdk and have it running in the debugger on an  
>>> iPad.
>>>
>>> I'd like to turn this work over to a macports developer to be  
>>> vetted, completed and merged into the mainline build in an  
>>> appropriate fashion. I didn't know much about the internals of  
>>> macports and I didn't known anything about TCL. I figured out most  
>>> of the details as I went along. I'd like to turn the work over to  
>>> someone more qualified than I am in scripting langauges and  
>>> macports to complete the project.
>>>
>>> Is there any interest of the part of the macports developers in  
>>> accepting the work?
>>>
>>>
>>> NOTE: I did not use your version control system. I modified the  
>>> scripts in place after the system was installed. I can provide a  
>>> list of the files I changed and some explanation on what was  
>>> changed. I basically attempted to piggy-back target/host support  
>>> on top of the existing variables without changing their function.  
>>> The approach is fine for proof-of-concept, but the the underlying  
>>> variables such as os.version and such should be modified and/or  
>>> renamed to make the design cleaner. However, that task should be  
>>> done by someone who knows the internals very well and who knows  
>>> how the changes will affect existing ports.  I did have to modify  
>>> the ICU port file because it has some custom steps that need to be  
>>> done before its configure script will succeed for a cross-compile.  
>>> I also had to make an adjustment because of the lack of tzfile.h  
>>> in the target OS.
>>
>> Hi James --
>>
>> Sounds interesting -- could you post your changes as a unified diff?
>>
>> Cheers,
>> Landon
>
> _______________________________________________
> macports-users mailing list
> macports-users at lists.macosforge.org
> http://lists.macosforge.org/mailman/listinfo.cgi/macports-users