Re: [Iaxclient-devel] libiax2/iaxclient format negotiation

[Mikael M. <mik...@tj...>]

Steve Kann wrote:
> Mikael Magnusson wrote:
> 
>> Steve Kann wrote:
>>
>>>
>> ...
>>
>>>
>>> libiax2 API:
>>>
>>> Add params to iax_call to set format and capability.
>>> Add param to iax_accept to tell it what format we want.
>>>
>>> -extern int iax_call(struct iax_session *session, char *cidnum, char 
>>> *cidname, char *ich, char *lang, int wait);
>>> -extern int iax_accept(struct iax_session *session);
>>> +extern int iax_call(struct iax_session *session, char *cidnum, char 
>>> *cidname, char *ich, char *lang, int wait, int format, int capability);
>>> +extern int iax_accept(struct iax_session *session, int format);
>>>
>>> Let me know what you think..
>>>
>>> I've only tested this on one outgoing calls, so a review and testing 
>>> is in order..
>>
>>
>> ...
>>
>>>
>>
>> Will you send a patch to the libiax2 author? And will it be backward 
>> compatible, or will it require source modifications?
> 
> 
> 
> I have sent in a patch to the tracker, which includes all of my libiax2 
> changes up until current iaxclient CVS.  
> http://bugs.digium.com/bug_view_page.php?bug_id=0002866
> 

I have browsed the patch. Isn't align.c unnecessary complicated? Doesn't 
it work to use plain memcpy instead? I don't have access to a SPARC to 
test with, but the following may work. (Btw SOLARIS also runs on little 
endian systems also I think, but align.c is for big endian only)

unsigned short get_align16(unsigned short *src)
{
     unsigned short var;

     memcpy(&var, src, sizeof(unsigned short));
     return var;
}

unsigned int get_align32(unsigned int *src)
{
     unsigned int var;

     memcpy(&var, src, sizeof(unsigned int));
     return var;
}

int set_align32(unsigned int *dst, unsigned int src)
{
     memcpy(dst, &src, sizeof(unsigned int));
}

Regards,
Mikael Magnusson