Building for both i386 and ppc on Intel Mac

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Alexei and Benny,

I did test this on a PPC, it appeared to be fully functional, at least  
with our application. Did your client do a full recompile (make  
realclean)? Also, I don't think I tried more than one codec (speex),  
so could you try out a number of codecs?

@Benny: I just noticed this, thanks for propagating the patches! :)

Ruud Klaver
AG Projects

On 21 Mar 2009, at 20:16, Alexei Kuznetsov wrote:

> Hello Benny and Ruud,
>
> Have you tested Ruud's patches on PowerPC Macs? One of our users
> claims that he still has distorted sound problem on PPC. I had applied
> patches myself on 1.0.1 release before they hit the repository on 1.1
> branch.
>
> Alexei
>
> On Fri, Mar 13, 2009 at 3:32 PM, Benny Prijono <bennylp at teluu.com>  
> wrote:
>> On Wed, Feb 18, 2009 at 2:17 PM, Ruud Klaver <ruud at ag-projects.com>  
>> wrote:
>>>
>>> Hi Benny,
>>>
>>> On 13 Feb 2009, at 18:22, Benny Prijono wrote:
>>>
>>>> Thanks for the patch Ruud. The crc32 is used only by STUN related  
>>>> things
>>>> I think (so this includes TURN and ICE), and there are two sets  
>>>> of tables
>>>> for each endianness. I assume the table based crc32  
>>>> implementation works
>>>> fine when the library is built natively? If it doesn't then  
>>>> probably the
>>>> table is wrong, but if it does then I don't know why.
>>>>
>>>> cheers
>>>>  Benny
>>>
>>> It turns out the table was not wrong after all, the final result  
>>> just
>>> needed to be byteswapped on big endian. I've included a simple  
>>> patch which
>>> does this. pjlib-util-test works on ppc now.
>>>
>>
>> It's been so long, I hope you still remember this. :)
>>
>> I've just applied your two patches:
>>  - http://trac.pjsip.org/repos/ticket/749 for the universal binary
>> endianness detectino
>>  - http://trac.pjsip.org/repos/ticket/751 for the CRC32 big endian  
>> problem
>>
>> Re: crc32, probably the proper fix is to remove the big endian  
>> detection at
>> all, and just use one table for both! But I don't have a big endian  
>> machine
>> to play with, so I guess it's fine as it is.
>>
>> I also added your info about building universal binary in the FAQ:
>> http://trac.pjsip.org/repos/wiki/FAQ#osx-universal
>>
>> Thanks again for all Ruud.
>>
>> cheers
>>  Benny




[Index of Archives]     [Asterisk Users]     [Asterisk App Development]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Security]     [Bugtraq]     [Linux]     [Linux OMAP]     [Linux MIPS]     [Linux API]
  Powered by Linux