challenge: implement the max function

[Andrei Alexandrescu (See Website For Email)]

Frits van Bommel wrote:
> Lionello Lunesu wrote:
>> "Frits van Bommel" <fvbommel at REMwOVExCAPSs.nl> wrote in message 
>> news:ep0j0a$1av2$1 at digitaldaemon.com...
>>> Lionello Lunesu wrote:
>>>> "Andrei Alexandrescu (See Website For Email)" 
>>>> <SeeWebsiteForEmail at erdani.org> wrote in message 
>>>> news:45B3B243.2010103 at erdani.org...
>>>>> Here's a simple challenge: implement the max function. Requirements:
>>>>>
>>>>> a) generic
>>>>> b) efficient
>>>>> c) preserve lvalueness when possible such that one can write e.g.
>>>>>
>>>>> max(arr[0], arr[1]) *= 0.9;
>>>> ...when would that be useful? And, what happens if arr[0]==arr[1] ?
>>> Can't immediately think up an example of usefulness, but what 
>>> presumably happens if they're equal is that one of them gets reduced 
>>> by 10% (and rounded to the accuracy of typeof(arr[0])).
>>
>> Which one? :)
> 
> That would be implementation-dependent.
> If you need it to be a specific one, that's an additional requirement. 
> So implement the function yourself ;).
> But perhaps it doesn't matter?
> Or maybe the caller knows it just can't happen?

For conformity purposes, let's just say that in case the two are equal, 
the first should be picked.

Andrei