Re: [boost] [TR1 Review] TR1 Formal Review Beginning

From: Alexander Nasonov (alnsn-boost_at_[hidden])
Date: 2005-09-28 09:43:34

> Beman Dawes wrote:
> ...

I spent a couple of hours reading and commenting documentation.
I also looked into boost/tr/functional.hpp and boost/tr1/memory.hpp.
My review covers documentation only.

Header Include Style
--------------------

> #include <boost/tr1/memory.hpp>

It's fine with me even in spite of one small inconsistency. Why the path starts
with boost although the stuff is in namespace std?

Reference Wrappers
------------------
Smart Pointers
--------------

No dot in the end. It's inconsistent with other subsections
of 'TR1 By Subject' section.

Random Number Generators and Distributions.
-------------------------------------------
Mathematical Special Functions.
-------------------------------

My knowledge of these libraries is very limited so I skip them.

Tuples.
-------

> 6.1 Tuple types Containers 80
This sentense doesn't begin with C++ comments and sounds rather strange to me.

> Interoperability withy std::pair is only partially supported.

1. Someone already noticed a typo (withy) and 
2. "partially supported" doesn't give any clue what's not supported.

Hash Function Objects.
----------------------

1. No reference to Boost.Hash documentation.
2. It's not documented that boost version of std::tr1::hash
   is defined on a wider set of types then required by 6.3.3/1.
   It doesn't have to be documented but boost::hash has some side
   effects which worth mentioning in "Standard Conformity" section.
   For example, boost::hash is valid for enums even if hash_value
   is not specialized (BTW, why enums are not listed in 6.3.3?).

I vote to accept Boost.TR1.
--
Alexander Nasonov