Re: [boost] [http] Formal review of Boost.Http

Subject: Re: [boost] [http] Formal review of Boost.Http
From: Robert Ramey (ramey_at_[hidden])
Date: 2015-08-07 20:24:16

On 8/7/15 3:28 PM, Vinícius dos Santos Oliveira wrote:

> Delay acceptance of the library will only delay its usage.

I would disagree here.  There is no reason people can't start using it 
now.  In fact, it's much easier to get a library accepted if people have 
already started to depend on it.

> 2. Before presenting for review
>> normal library builds cannot use cmake OR should be header only like
>> Hana.

I don't think that there is a requirement that boost build be supported 
in order to review a library.
>
> CMake will be replaced by Boost.Build before any integration.

I don't think it's necessary to remove CMake support.  No reason you 
can't leave it in and have both CMake and Boost Build support.

>
> Until you address a minimum of items 1 and 4, I am very sorry but I
>> must vote for rejection.

Picking a nit here. The review manager is under no obligation to weigh 
"votes" equally.  So I think it would be better if we used the word 
"recommendation" rather than vote.

> * You're still missing CI testing with valgrind, thread sanitiser,
>> coveralls.io coverage testing, etc etc all the stuff from
>> https://svn.boost.org/trac/boost/wiki/BestPracticeHandbook.

The items listed in 
https://svn.boost.org/trac/boost/wiki/BestPracticeHandbook do not 
represent any official boost policy nor have they been discussed such 
that they represent any consensus.  So though they may be relevant to 
Nail's recommendation, they may not be relevant to anyone else's.

Robert Ramey