RE: [boost] Re: Re: Formal review queue

From: Matthew Hurd (matt_at_[hidden])
Date: 2004-02-25 14:46:07

> On Behalf Of Michael Glassford
> Subject: [boost] Re: Re: Formal review queue
> 
> Were there any changes that particularly concern you?
I'm not too concerned about much and most things should be resolvable with a
little email conversation via the list.

Trivia like the name of rw_mutex, ensuring the orthogonality of the
interface, making sure a little boolean conversion sneaks in for a lock.
Things like that.

I'm not clever enough to get beyond trivia...

Maybe splitting boost::thread into separate namespaces.  Fundamentally
mutexes, locks and some other things might have nothing to do with threads,
the thread effect might be implicit, but another namespace also seems
overkill.  I worry that concurrency and synchronisation do not mean explicit
threading.  

It does raise the issue of the size and scale of library changes as it seems
that library approval indicates author approval to make future changes and
though this seems to have worked very well so far, it will fail at some
stage.  Nothing lasts for ever (homily about my extraordinary wife omitted).

I think you, Mr Michael G, is doing a lovely job of tackling boost::thread
and I do appreciate your efforts.

Regards,

Matt Hurd.