Re: RFC: status icon support

[David Kastrup <dak@gnu.org>]

Andreas Schwab <schwab@suse.de> writes:

> David Kastrup <dak@gnu.org> writes:
>
>> Andreas Schwab <schwab@suse.de> writes:
>>
>>> David Kastrup <dak@gnu.org> writes:
>>>
>>>> Richard Stallman <rms@gnu.org> writes:
>>>>
>>>>>     Please no K&R in new code.  
>>>>>
>>>>> We have no policy against K&R style.  I recently accepted non-K&R
>>>>> function definitions in Emacs sources, but I still do not particularly
>>>>> like it.  K&R style is easier to read anyway.
>>>>
>>>> It stops the compiler from doing type checking and type conversions, for
>>>> one thing.
>>>
>>> Unless you add prototypes.
>>
>> In which case you have
>> a) undefined behavior (the compiler need not check whether the prototype
>>    matches the function declaration).
>
> There is nothing undefined in old-style function defintions.

There certainly is.  If they don't match the prototype, the compiler is
free to do whatever it thinks correct.  It is not required to generate
an error.

>> b) not K&R compatible C, so where is the point?
>
> The Emacs sources use macros that keep prototypes from being visible
> from obsolte compilers.

We stopped mandating this.

> What's your point?

Where is the point in combining a prototype that won't compile under K&R
compilers with a K&R definition (which won't compile under C++ or newer
standard compilers)?

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum