Re: FilterCompressedFiles

[Debajyoti Bera <dbera web gmail com>]

Hi,

> > > I've been playing a bit with beagle creating an small filter, this
> > > look into some compressed files, and index the file names inside the
> > > compressed file.

Good start. There is a bug open in bugzilla about indexing rar and 7z 
archives. Your filter would be useful for that.

> > Does this mean that your filter and the filters from FilterArchive.cs
> > both run over all archives when indexing ?
>
> No, my filter does not run over all archives, it only index the file
> names and paths that are inside the compressed file.
>
> FilterArchive.cs index the file names too, and add some properties
> too, like comments, mimetype, ...
>
> The difference on both filters is that FilterArchive.cs works only
> over the archives supported by ICSharpCode.SharpZipLib (currently:
> zip, gzip, bzip2 and tar), while my filter could work with any format
> supported by any external tool, ie: unzip, unace, unrar, tar, p7zip,

Ideally we would like to index not only the names and paths but also the 
contents. Whenever we cannot index the contents, we at least filter the names 
and paths. Since FilterArchive already indexes the contents of tar, zip etc. 
I believe your filter can be used to handle non-SharpZipLib supported 
formats.

> Will be interesting to create a filter that reads a configuration file
> that, and apply the filter to the mimetypes included in the config,
> extracting the information using the command-line tools and regular
> expressions described in the config?

Upon second thoughts, if FilterCompressedFiles does not do anything other than 
extracting names and paths, then it can be already done using External 
Filter. e.g. I use the following configuration to index the names of my rar 
archives (requires the rar executable from rarlabs.com):

<filter>
  <mimetype>application/x-rar</mimetype>
  <extension>.rar</extension>
  <command>/mnt/extra/files/rar/rar</command>
  <arguments>lb %s</arguments>
</filter>

I am sure similar thing can be done with the other unsupported archive 
formats. A separate filter would only make sense if it can extract 
interesting properties of the archive or extract the archive contents.

- dBera

-- 
-----------------------------------------------------
Debajyoti Bera @ http://dtecht.blogspot.com
beagle / KDE fan
Mandriva / Inspiron-1100 user