Re: [PATCH trivial] include/linux/gfp.h: Improve the coding styles

[Chen Gang]

On 2/28/16 07:14, Jiri Kosina wrote:
> On Sat, 27 Feb 2016, Chen Gang wrote:
> 
>>> Mel, as an MM developer, has already NACK'ed the patch, which means
>>> you should not send the patch to **any** upstream maintainer for
>>> inclusion.
>>
>> I don't think I "should not ...". I only care about correctness and
>> contribution, I don't care about any members ideas and their thinking.
>> When we have different ideas or thinking, we need discuss.
> 
> If by "discuss" you mean "30+ email thread about where to put a line 
> break", please drop me from CC next time this discussion is going to 
> happen. Thanks.
> 

Excuse me, when I sent this patch, I did not know who I shall send to, I
have to reference to "./scripts/get_maintainer.pl".

If any members have no time to care about it (every members' time are
really expensive), please let me know (can reply directly).

Thanks.

>> For common shared header files, for me, we should really take more care
>> about the coding styles.
>>
>>  - If the common shared header files don't care about the coding styles,
>>    I guess any body files will have much more excuses for "do not care
>>    about coding styles".
>>
>>  - That means our kernel whole source files need not care about coding
>>    styles at all!!
>>
>>  - It is really really VERY BAD!!
>>
>> If someone only dislike me to send the related patches, I suggest: Let
>> another member(s) "run checkpatch -file" on the whole "./include" sub-
>> directory, and fix all coding styles issues.
> 
> Which is exactly what you shouldn't do.
> 

For me, I also guess, I am not the suitable member to do that (in fact,
I dislike to do like that - "run checkpath -file" on "./include").

> The ultimate goal of the Linux kernel is not 100% strict complicance to 
> the CodingStyle document no matter what. The ultimate goal is to have a 
> kernel that is under control. By polluting git blame, you are taking on 
> aspect of the "under control" away.
> 

Yes, the ultimate goal of CodingStyle is to have a kernel that is under
control.

For me, most of files in "./include" are common, simple, and shared
files, which are not quite related with code analyzing (e.g. git log -p,
or git blame), but they are read by others in most times. Is it correct?


> Common sense needs to be used; horribly terrible coding style needs to be 
> fixed, sure. Is 82-characters long line horribly terrible coding style? 
> No, it's not.
> 

For me, what you said above have effect on body files (in kernel, at
least, more than 95% source files are body files, I guess).

But in "./include", most of files are the interface inside and outside
of our kernel, we need take more care about their coding styles.

I often use vertical split window in vim in full screen mode to reading
source code, when I read c source files, I often split window vertically
for the related header files as reference.


Thanks.
-- 
Chen Gang (éå)

Managing Natural Environments is the Duty of Human Beings.