[FFmpeg-devel] [PATCH] make libavcodec use bytestream functions

[Benoit Fouet]

Ramiro Ribeiro Polla wrote:
> M?ns Rullg?rd wrote:
>> Ramiro Ribeiro Polla wrote:
>>  
>>> Ramiro Ribeiro Polla wrote:
>>>    
>>>> M?ns Rullg?rd wrote:
>>>>      
>>>>> Jindrich Makovicka wrote:
>>>>>
>>>>>        
>>>>>> On 5/10/07, Jindrich Makovicka <makovick at gmail.com> wrote:
>>>>>>
>>>>>>          
>>>>>>> On 5/10/07, M?ns Rullg?rd <mans at mansr.com> wrote:
>>>>>>>
>>>>>>>            
>>>>>>>> Jindrich Makovicka <makovick at gmail.com> writes:
>>>>>>>>
>>>>>>>>              
>>>>>>>>> intreadwrite.h needs fixing to compile here.
>>>>>>>>>
>>>>>>>>>                 
>>>>>>>> OK.
>>>>>>>>
>>>>>>>> Out of interest, what compiler are you using, and what is the
>>>>>>>> error
>>>>>>>> message?
>>>>>>>>
>>>>>>>>               
>>>>>>> i am away from the machine now, but it's gcc from Debian Sid
>>>>>>> (probably
>>>>>>> 4.1.2) complaining about the semicolon after the closing brace.
>>>>>>>
>>>>>>>             
>>>>>> more precisely, it complained about this piece of code:
>>>>>>
>>>>>>         if (le) AV_WL16(*dst, v);
>>>>>>         else    AV_WB16(*dst, v);
>>>>>>
>>>>>>           
>>>>> Ah, that explains it.  That bit is invalid C with the braces as they
>>>>> are now.
>>>>> Where is that code?
>>>>>         
>>>> Should I fix this in pcm.c with attached patch or intreadwrite.h?
>>>>
>>>>       
>>> Fixed in pcm.c, since it seems more correct.
>>>     
>>
>> It was not correct.  The header is still broken, and anything that looks
>> like valid code should remain that way after macro expansion.
>>
>>  
>>> The ST macros aren't enclosed in {}s, so that's why it compiled fine
>>> for
>>> me. When more than one byte is written to in intreadwrite.h, {}s are
>>> used, so the ; after them screwed up the if statement.
>>>
>>> Attached patch prevents this error from being silently ignored
>>> again. It
>>> makes all AV_W macros look the same as {}s are concerned.
>>>     
>>
>> No, very much no.  Your first patch was good.  Apply it.
>>
>>   
> Do you prefer this patch? It looked uglier to me, so that's why I
> preferred fixing pcm.c.
> But with your explanation, it sounds saner.
>
> After intreadwrite.h gets fixed, should I revert the temporary fix I
> made on pcm.c (which wasn't the best solution)?
>

i think M?ns was talking about Jindrich fix, no ?
with the do{...}while(0) stuff

Ben
-- 
Purple Labs S.A.
www.purplelabs.com