httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Greg Stein <gst...@lyra.org>
Subject Re: bucket->copy()
Date Thu, 09 Nov 2000 01:48:28 GMT
If you're up for doing a copy(), then go for it. I'm happy to get back to
SVN work :-) The byterange filter also needs to use split (which also
implies implementing split in the rest of the buckets).

Is there any part you'd rather that I do, or would you like to tackle them?

Cheers,
-g

On Wed, Nov 08, 2000 at 05:33:26PM -0800, rbb@covalent.net wrote:
> 
> Yeah, I was calling this a duplicate function in all of my conversations
> with people today.  If nobody beats me to it, I'll be doing it in a few
> days.
> 
> Ryan 
> 
> On Wed, 8 Nov 2000, Greg Stein wrote:
> 
> > I'd like to introduce a bucket->copy() function into our bucket design.
> > 
> > Motivation: consider the byterange filter. It needs to select subsets of a
> > brigade and send them to the client. We can remove the content from the
> > original brigade in a copy-free manner, but this is a destructive
> > modification to the original brigade -- we may need it for multiple ranges.
> > 
> > This can also eliminate the big alloc in the filter. Rather than copy all
> > the data from the original brigade into a memory buffer(!), we can copy
> > portions of the buckets and insert those into the output brigade.
> > 
> > Cool part: all of the refcounted buckets can support copy with *LITTLE*
> > effort. These included: HEAP, POOL, and MMAP. FILE should be, but it isn't
> > code that way right now. The TRANSIENT and IMMORTAL are butt-simple because
> > they would just refer to the same data. EOS and FLUSH are cake, of course.
> > 
> > PIPE and SOCKET are the tougher ones since they must read first, then copy
> > the result. But since they read into a HEAP... they are easily copied.
> > 
> > 
> > Given a copy operation, we should be able to do byterange handling without
> > any data copies.
> > 
> > Cheers,
> > -g
> > 
> > -- 
> > Greg Stein, http://www.lyra.org/
> > 
> > 
> 
> 
> _______________________________________________________________________________
> Ryan Bloom                        	rbb@apache.org
> 406 29th St.
> San Francisco, CA 94131
> -------------------------------------------------------------------------------

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

Mime
View raw message