Mailing-List: contact dev-help@apr.apache.org; run by ezmlm
Date: Mon, 11 Dec 2000 15:46:19 -0800
From: Greg Stein <gstein@lyra.org>
To: dev@apr.apache.org
Subject: brigade/bucket splitting (was: Re: cvs commit: apr-util STATUS)x
Message-ID: <20001211154618.U7732@lyra.org>
Mail-Followup-To: dev@apr.apache.org
References: <003201c06212$a3cd7ba0$92c0b0d0@roweclan.net>
 <Pine.LNX.4.21.0012091313480.24264-100000@koj.rkbloom.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2i
In-Reply-To: <Pine.LNX.4.21.0012091313480.24264-100000@koj.rkbloom.net>;
 from rbb@covalent.net on Sat, Dec 09, 2000 at 01:14:57PM -0800

On Sat, Dec 09, 2000 at 01:14:57PM -0800, rbb@covalent.net wrote:
>...
> I'll buy this.  I can see this being somewhat useful.  Part of me thinks
> this should be implemented as a brigade_seek function.  The calling code
> would then be responsible for calling ap_bucket_split and
> ap_brigade_split, but I am not sure if I agree with that or not yet.

I've been thinking of a bucket split at a specified point. That allows you
to copy the buckets before/after the split point.

This also means we're copying *a* bucket, rather than needing to pass
start/end values into a bucket copy.

If we go back to my original email, and we have a brigade such as:

    B1 -> B2 -> B3/4 -> B5 -> B6/7 -> B8

And we want to copy from [4..6], then we split at the correct point within
bucket B3/4 and B6/7. That gives us this brigade:

    B1 -> B2 -> B3 -> B4 -> B5 -> B6 -> B7 -> B8

Now we can copy B4..B6.

The reason that I asked for a bucket return value was so that we could split
the brigade at point P3/4, and get a pointer to B4 back (the start of the
copy). We split at point P6/7 and get P7 back (the bucket after the last to
copy).

Cheers,
-g

-- 
Greg Stein, http://www.lyra.org/