[bind10-dev] dependency on boost runtime libraries

[JINMEI Tatuya / 神明達哉]

At Wed, 24 Mar 2010 20:59:48 +0100,
Jelte Jansen <jelte at isc.org> wrote:

> > - declare that we don't support (or guarantee) this type of
> >   compatibility
> 
> Can we? Does boost have any form of stability guarantee?

In this bullet I meant "declare that we do neither support nor
guarantee this type of compatibility", that is, this "solution" is to
ignore difficult cases.

> > - hide any non-builtin types from public interfaces (or provide
> >   equivalent interfaces only using builtin types)
> 
> IMHO, this is the path we should take (and then we still have to decide on
> header-only or not and whether to supply those headers :) )

I agree this is ideal in terms of wider portability and
deployability.  A big con with this approach is that it will make the
implementation more complicated and probably more buggy.  For example,
by using plain old pointers instead of shared pointer objects, we'll
increase the risk of memory leak, use of dangling pointers, etc.
Exception handling will also be much more complicated.

> > Each has its pros and cons.  We'll eventually have to make a decision,
> > but it'll probably a bit later phase of the development when we have
> > more complete set of features.
> 
> Well, I'm afraid that if we don't deal with this sooner rather than later, the
> problem creeps into every corner...

I generally agree; what I meant is that even if we decide something
we'll see other difficulty as we complete missing features.  But we
should probably make a general guideline on this point at an earlier
stage of the year2 development (a topic of the next f2f?).

---
JINMEI, Tatuya