Boost logo

Boost :

Subject: Re: [boost] [modularization] proposal and poll
From: Niall Douglas (s_sourceforge_at_[hidden])
Date: 2014-05-30 07:37:06

On 30 May 2014 at 12:04, John Maddock wrote:

> > Does this make sense? I am specifically advocating a structural
> > break, a clean start. I think working on a fresh, empty Boost will
> > get people to volunteer to give up family time to write Boost code
> > again.
> All I see from this idea is a requirement to spend even more time on
> Boost. Just my 2c yours, John.

It should be fun again to work across Boost instead of within silos

So why do I keep saying end of this year?

1. I need to try removing Boost as a dependency from all my own code
first to figure out what really is the minimum needed. I'm blocked on
expected<> for this right now, I need that to eliminate the need for

2. I need a meta test framework which lets you whatever underlying
test framework. This lets me swap out Boost.Test for any other test
framework, I really don't care which so long as it outputs Jenkins CI

3. Keep building out my CI based quality assessment framework for the
wider AFIO project, for which I need at least to write:

  a. A clang AST parser which warns on things like move constructors
without noexcept and other things I keep forgetting in my code etc.

  b. A clang source rewriter which does a live Boost to STL source
conversion. This will mirror the minimum of Boost I need into a C++
14 edition, per commit to original Boost.

4. I'm thinking of all this as being a great showcase for my proposed
graph database. See my C++ Now presentation on what I proposed as a
future "as if all header only" C++ build system.

I'm basically writing all this for my own purposes, but if others
find it useful, they can hop onboard. I hope to restart work on all
this after June 15th, health permitting. It's been a very long hiatus
these last many months sadly.


ned Productions Limited Consulting

Boost list run by bdawes at, gregod at, cpdaniel at, john at