Re: kernel.h: add ARRAY_AND_SIZE() macro to complement ARRAY_SIZE().

[Cyrill Gorcunov <gorcunov@gmail.com>]

[Christer Weinigel - Sat, Sep 20, 2008 at 03:05:29PM +0200]
> Alexey Dobriyan wrote:
>> On Fri, Sep 19, 2008 at 08:28:45AM -0700, Russ Dill wrote:
>>> My vote is for ARRAY_AND_SIZE to spread far and wide across the land.
>>> ARRAY_SIZE is already very safe, as it has a __must_be_array macro
>>> built in. So ARRAY_AND_SIZE is even safer, as it prevents you from
>>> mixing up two different arrays. It also reduces line length and makes
>>> driver and device (usually platform_device) registration code easier
>>> to read.
>>
>> It also spreads ARRAY_SIZE misnaming futher.
>
> You still haven't explained what's misnamed about it, nor suggested a  
> better name.
>
>> It introduces one more core macro and quite pointless one. I can't
>> personally recall a single bug where sizeof() was taken from another
>> array.
>
> You haven't written a lot of machine definitions then.  When adding  
> platform devices for an embedded platform one has to write a lot of  
> boilerplate like this:
>
>     platform_add_devices(n30_devices, ARRAY_SIZE(n30_devices));
>
> and it is much too easy to copy paste that line and miss one of the  
> references.
>
>> It creates interesting confusion point: ARRAY_AND_SIZE is about array
>> and it's size. What ARRAY_SIZE is about then?
>
> ARRAY_AND_SIZE -> (An) array and (its) size
>
> ARRAY_SIZE -> (The) array size
>
> Sure, you could write ARRAY_AND_ITS_SIZE, but would that really make  
> anyone happy?  Cobol went out of fashion a long time ago.
>
>   /Christer
>

Christer, _I_ was complaining not about naming
but about hiding function arguments. I suppose
it's better to define some inline wrapper for
platform_add_devices then use such a macro.

If you google a bit you may find that I was asking
someday why don't we define alias for memset when
we just it as _clearing_ routine ie memset(x, 0, sizeof(x)).

		- Cyrill -