Re: [dpdk-dev] [PATCH v3 00/40] Fix build on gcc8 and various bugs

[Thomas Monjalon <thomas@monjalon.net>]

Hi,

11/05/2018 02:29, Andy Green:
> 
> On 05/10/2018 11:01 PM, Stephen Hemminger wrote:
> > On Thu, 10 May 2018 20:13:31 +0800
> > Andy Green <andy@warmcat.com> wrote:
> > 
> >> I appreciate the reply.
> >>
> >> But why bother having a subject line at all if it is going to be
> >> mechanically enforced that nothing in it is allowed to be "useful"?
> >> That really doesn't make sense does it.
> > 
> > It was done because there were lots of clueless patches showing
> > up on the driver development list which had useless subject
> > lines.

Yes, the title should help to quickly identify the scope of the commits.
Usually, giving variable names, function names, etc are not very useful.
It is very common that some developers say "fixing function X" instead
of describing the fixed behaviour: "fixing feature Y".

> The "cure" is worse than the disease...

It is a tool showing some warnings.
We must be smart when using this tool (as any other tools)
and consider that some warnings are false positive.

>   - I can mention, eg, that something changed to an int.  But a size_t 
> or my_type_t?  I am not allowed to mention it even if that is the whole 
> reason for the patch.

Of course you can use size_t in the title if it is relevant.
Sometimes, changing a type is fixing a feature, so better to name the feature.

>   - I can mention most libc apis, but not those that happen to have an 
> underscore, eg, timerfd_create(), even if that was the focus of the patch.
> 
>   - Any kind of manifest constant like MY_CONSTANT: illegal to mention, 
> even if the patch's job is change MY_CONSTANT to, say, 5.  What should I 
> entitle that patch?  "lib: change something to 5"?  "lib: change 
> MY.CONSTANT to 5"?
> 
>   - I can mention most filenames or paths, eg, down /proc, or myfile.c. 
> But not if the filepath happens to contain an underscore.  Even if the 
> effect of the patch is to migrate stuff from myfile.c to my_files/
> 
> The results are arbitrary... please consider removing this now it has 
> been in place a while and made its original point.

I hope you got my point that the title should be a high level description
of the scope or goal of the patch.
The details are inside the commit log.
But using underscore is accepted in the title if relevant.

Every strong rules are stupids, that's why we have only guidelines
and we are flexible.