[boost] Re: Re: : [chrono] v0.4.5 Documentation up Aate + warnings +removal +>ug fixes

Subject: [boost] Re: Re: : [chrono] v0.4.5 Documentation up Aate + warnings +removal +>ug fixes
From: vicente.botet (vicente.botet_at_[hidden])
Date: 2010-08-15 14:59:32

----- Original Message ----- 
From: "vicente.botet" <vicente.botet_at_[hidden]>
To: <boost_at_[hidden]>
Sent: Friday, August 13, 2010 6:50 PM
Subject: [boost]Re: : [chrono] v0.4.5 Documentation update + warnings +removal +>ug fixes

> 
> ----- Original Message ----- 
> From: "Tan, Tom (Shanghai)" <TTan_at_[hidden]>
> To: <boost_at_[hidden]>
> Sent: Friday, August 13, 2010 10:25 AM
> Subject: Re: [boost] : [chrono] v0.4.5 Documentation update + warnings +removal + bug fixes
> 
> 
>> >Date: Tue, 10 Aug 2010 20:00:15 +0200
>>>From: "vicente.botet" <vicente.botet_at_[hidden]>
>>>Subject: [boost] [chrono] v0.4.5 Documentation update + warnings
>> removal + bug fixes
>>>
>>>Waiting for the review I plan to add yet a lightweight stopwatch, as
>> suggested by Tom Tan on a personal mail, 
>>>which stores a reference to a generalization of the Clock duration or
>> an accumulator of the Clock duration.
>>> I'll concentrate also in improving the documentation and the tests. 
>> Hi Vincente, 
>> Thanks for the feature. One more request, can you make this following
>> specialization work?
> 
> Could you elaborate a little bit?
> 
>>    //typedef stopclock_accumulator< boost::chrono::process_cpu_clock >
>> process_cpu_stopclock_accumulator;
>> It's commented out in your source code, I assume it does not compile
>> yet. 
> 
> Yes. It doesn't compile because boost::chrono::process_cpu_clock is not really a Clock. 
> This line is also commented:
>    //typedef stopclock< boost::chrono::process_cpu_clock > process_cpu_stopclock;
> 
>> In general, I'd like to see process_cpu_clock to be supported as
>> the first-class citizen like process_real_cpu_clock,
>> process_user_cpu_clock and so on.
> 
> Me too. Long time ago a tryed it, but I didn't reach to model it completly in terms of the Clock requirements. I will try again if find an idea on how to do it.

I have tried again, and what was related to the interface required by the accumulator and the formatter. I have added the multiplicative operators with two arguments (times,int) and the accumulator works now. The formatter stay a problem as I need a formatter that takes care of the tuple nature of process_cpu_clock  and the statistical results. 

Tom, what kind of statistical report do you expects in this case?

Best,
Vicente