Re: [PATCH 4/4] test_printf: test printf family at runtime

[Rasmus Villemoes <linux@rasmusvillemoes.dk>]

On Wed, Sep 30 2015, Andy Shevchenko <andriy.shevchenko@linux.intel.com> wrote:

> It's still better to distinguish what function does by names
> test vs. __test confuses me.

Think of test() as the ordinary "public" interface and __test() as
"private, but we expose it because you might need it, if you know what
you're doing". Having two related functions named like this is a
perfectly normal pattern in the kernel (even if one isn't necessarily
implemented in terms of the other); examples are
legio. krealloc/__krealloc (the latter is not called krealloc_nofree),
fls/__fls (not fls_I_know_word_is_nonzero, thank $deity), ...

>> >Maybe commentary delimiter here and above where you have double
>> >empty line.
>> 
>> And say what? I can avoid double empty lines if they bother you.
>
> So, what was the reason to add them in the first place?

Just so we could have this interesting little chat. I don't think it was
intentional, but don't worry, they're gone.

>> > > +MODULE_AUTHOR("Rasmus Villemoes <linux@rasmusvillemoes.dk>");
>> > > +MODULE_LICENSE("GPL");
>> > 
>> > GPL or ?..
>> 
>> Honestly, I don't really care. Would you like BSD/GPL or what? I just
>> copied from the majority of MODULE_LICENSE() instances.
>
> You are the author, your choice. I'm okay with any applicable type.

Then I'll stick to GPL, even if that's more cargo-culting than
deliberate informed choice.

Rasmus