RE: mod_disk_cache backward compatibility? Re: [PATCH] [UPDATED] Re: Cached response: 304 send as 200

[Brian Pane <br...@apache.org>]

For now, I've left the status line in the cache header; the
reader code just reads that line and ignores it.  (Thanks to
Kris for the patch.)

Is there any reason why the header and the content for a cache
entry are stored in two separate files?  Concatenating them into
the same file would allow for faster reads, faster cache updates,
and fewer inodes per cache entry.

Brian


On Fri, 2002-09-13 at 13:22, Bill Stoddard wrote:
> 
> 
> > -----Original Message-----
> > From: William A. Rowe, Jr. [mailto:wrowe@rowe-clan.net]
> > Sent: Friday, September 13, 2002 3:52 PM
> > To: dev@httpd.apache.org
> > Cc: dev@httpd.apache.org
> > Subject: RE: mod_disk_cache backward compatibility? Re: [PATCH]
> > [UPDATED] Re: Cached response: 304 send as 200
> >
> >
> > At 02:42 PM 9/13/2002, Bill Stoddard wrote:
> > > > Kris Verbeeck wrote:
> > > >
> > > > Thanks!  One more issue before I commit this...
> > > > The change will require that users clear their file
> > > > caches.  Can any of the mod_cache hackers give me a
> > > > second opinion on this?  Since mod_file_cache is still
> > > > in experimental state, it reasonable to just make the
> > > > change and document the need to clear one's file cache
> > > > when installing 2.0.4.  Or do we need to provide some
> > > > automated checking for this in mod_disk_cache (like
> > > > adding a magic number and format version number in
> > > > the file header)?
> > >
> > >Document the change and inform folks to clear their file cache.
> >
> > Folks don't read the docs.  We know it, just look at the bugzilla
> > database.  A magic number is much more appropriate, and will
> > make it possible to update or read old formats in the future.
> >
> > Bill
> 
> I am really not keen on crufting up the codebase to support backward
> compatability with a module that was unusable up until 2.0.40 and is
> experimental in 2.0.40.
> 
> Bill

RE: mod_disk_cache backward compatibility? Re: [PATCH] [UPDATED]Re: Cached response: 304 send as 200

[Bill Stoddard <bi...@wstoddard.com>]

> For now, I've left the status line in the cache header; the
> reader code just reads that line and ignores it.  (Thanks to
> Kris for the patch.)
> 
> Is there any reason why the header and the content for a cache
> entry are stored in two separate files?  
Not really.

> Concatenating them into
> the same file would allow for faster reads, faster cache updates,
> and fewer inodes per cache entry.

+1

Bill