[Xapian-discuss] zip/rar support

[Olly Betts]

On Sun, Aug 27, 2006 at 04:10:36PM +0100, James Aylett wrote:
> On Sun, Aug 27, 2006 at 02:08:30PM +0200, Reini Urban wrote:
> 
> > Yes. This extra script (a php file) checks the cache and sends the 
> > already extracted part to the browser, and if it not exists in the cache 
> > extracts it on the fly from the url to the browser.
> 
> I'm pretty certain I've seen an apache module that does this with the
> simplest possible URI layout:
> 
> /path/to/archive/path/to/packed/file

This looks like what you mean:

http://nobits.org/articles/2005/10/12/mod-unzip/

That webpage points out the zip files are indexed so accesses are
pretty fast - I suspect they're right and caching probably isn't worth
the effort for the contents of zip files unless your web server is CPU
bound and has I/O to burn.

> If not, with apache at least you could write it in PHP, Python or
> Perl, without needing to pre-unpack the archive. Adding in caching
> isn't terribly difficult.

I think this is a good approach, since it doesn't rely on a particular
script being installed in a particular location, and it degrades
reasonably gracefully if there's no handler installed to run mod_unzip
(or something equivalent) - in that case the web server will ignore the
extra path information and just download the whole zip file.  That's
much better than a 404 error.

We should be able to use the same approach for any "container" filetype
though the appropriate apache module or other handler script may not
already exist.

Cheers,
    Olly