On 22/05/17 14:23, Laurent Pinchart wrote:
> Hello Geert and Kieran,
>
> On Monday 22 May 2017 15:00:27 Geert Uytterhoeven wrote:
>> On Mon, May 22, 2017 at 2:52 PM, Kieran Bingham wrote:
>>> My only distaste there is having to then add the [i-1] index to the
>>> sg_tables.
>>>
>>> I have just experimented with:
>>>
>>> fail:
>>> for (; i-- != 0;) {
>>> struct sg_table *sgt = &rstate->sg_tables[i];
>>> ...
>>> }
>>>
>>> This performs the correct loops, with the correct indexes, but does the
>>> decrement in the condition offend coding styles ?
>>>
>>> If that's disliked even more I'll just apply your suggestion.
>>
>> You can still use "i-- > 0", which looks a little bit better IMHO.
>
> I'm fine with that option too.
>
Of course for(; X ;) is just while(X), which is also more readable ;)
And while (i-- > 0) simplifies cleanly to while (i--) which I'm sure is quite
readable.
I'll clean up and post the updated series including linux-media.
--
Kieran
