You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@apr.apache.org by Justin Erenkrantz <ju...@erenkrantz.com> on 2007/12/05 02:13:03 UTC

Re: APR 2.0 proposals

On Nov 23, 2007 2:37 AM, Joe Orton <jo...@redhat.com> wrote:
> ** Key proposal: "one tree, multiple libraries"

+1.

> ** Second (more controversial/radical?) proposal:
>
> Reduce the consolidated libapr library size by chucking out everything
> from apr-util which has been around for N years and is not used outside
> httpd.  Anything that is used only by httpd should be moved to the httpd
> tree, no point everyone else being burdened by it.  Also chuck out stuff
> which has *no* users at all.
>
> Impact: maybe gets rid of apr/random/, apr-util/buckets/,
> apr-util/ldap/, half of apr-util/misc?  (If/when httpd moves to the serf
> buckets model maybe buckets can go away completely?)

+1.  There's enough feedback to see what most of our apps actually use
and what's dead code or only used by httpd (hint: bucket brigades!).

> ** Third proposal:
>
> Goes without saying: break API/ABI with wild abandon (ish).

I know we've talked about it for a while: but what about punting pools
or replacing it with a vtable impl?  There's some nice code in APR
that I'd like to use every once in a while, but importing the whole
pool methodology gets a bit limiting.  For example, Serf tries to
primarily use an allocator scheme, but has to be tied to pools for
certain calls back into APR.  I shudder to think at what it'd be to
actually implement such a split...  =(  -- justin