Re: [Boost-users] C++ gurus - really dumb question regarding partial speciallization.

Subject: Re: [Boost-users] C++ gurus - really dumb question regarding partial speciallization.
From: Vicente J. Botet Escriba (vicente.botet_at_[hidden])
Date: 2012-12-20 02:43:25

Le 19/12/12 20:31, Jeremiah Willcock a écrit :
> On Wed, 19 Dec 2012, Robert Ramey wrote:
>
>> I've got the following code which fails to compile.  I'm really 
>> stumped by
>> this.
>> I've consulted the standard document on the subject and for the life 
>> of me
>> I can't see anything wrong with it.  It fails with the same message 
>> on vc
>> 9.0
>> as well as gcc 4.5.3.  It's so simple it can't be a compiler issue.
>>
>> I appoligize in advance for spaming the list here - but I'm really, 
>> really
>> stuck.
>>
>> Robert Ramey
>>
>> // primary template
>> template<
>>  class T,
>>  int MIN,
>>  int MAX
>>>
>> struct safe_range1{};
>>
>> // instantiate primary as a test
>> safe_range1<int, 0, 1> t1;
>>
>> // partial specialization # 1
>> template<
>>  int MIN,
>>  int MAX
>>>
>> struct safe_range1<int, MIN, MAX>{};
>>
>> // instantiate partial specialization - what the ???
>> // fails with "too few template parameters" !!!
>> safe_range1<0, 1> t2;
>>
>> // partial specialization # 2
>> template<class T>
>> struct safe_range1<T, 0, 1> {};
>>
>> // also fails with "too few template parameters" !!!
>> safe_range1<int> t3;
>
> Your template, and all of the partial specializations, have three 
> parameters and no defaults.  Partial specialization #1 just states 
> that when int is used as the first argument to safe_range1, it will 
> use that case, but that doesn't change the parameter type.
>
>
You could get something close using template alias (c++11), but of 
course you need to have a different name.

Vicente