Subject: Re: [boost] Boost reviews
From: Niall Douglas (s_sourceforge_at_[hidden])
Date: 2013-09-21 17:14:20
On 20 Sep 2013 at 22:42, Edward Diener wrote:
> > Did you not see my post
> > http://listarchives.boost.org/Archives/boost/2013/09/206187.php? I offered
> > to review manage someone's library in exchange for that person review
> > managing AFIO.
>
> I did see your post but my thoughts are:
>
> 1) It is not best if A is the review manager for B's library and B is
> the review manager for A's library. There is too much temptation to
> approve a library that way.
I agree. I can also see potential for heated arguments.
> 2) Understanding ASIO has proven to me to be a difficult task, albeit I
> have not spent much time studying it. As I understand it AFIO is an
> extension of ASIO for files.
Correct. ASIO is totally familiar if you're used to COM and WinRT
reactor style callback programming, otherwise I agree it's a real
mountain to climb.
And AFIO, I am told, is harder than ASIO due to the dependency
chaining facility which apparently wrecks people's heads.
> 3) With that last said I would be glad to be a review manager for a
> library. But My OP was prompted because I have seen little movement
> recently in the way of even calls for review managers to speed up the
> process of having libraries reviewed. Perhaps I just missed such a
> general request.
It's certainly a concern. Apart from directory monitoring, I expect
AFIO to be finished this coming week - docs and all. I've just
finished two days of performance tuning, and I have an AFIO
implementation blowing past a multithreaded implementaton by 25% in
warm cache, and a full 100% in cold cache.
It's basically only polish left to go - t's crossed, i's dotted etc.
Bugs notwithstanding of course, and there are many :(
Niall
-- Currently unemployed and looking for work. Work Portfolio: http://careers.stackoverflow.com/nialldouglas/