Re: [PLUGIN] vCards

[David Schneider <dnschneid gmail com>]

---------- Tue, 23 Mar 2010 20:39:31 +0100 ----------
Ulrik Sverdrup <ulrik sverdrup gmail com> wrote:

> 2010/3/6 David Schneider <dnschneid gmail com>:
> > This plugin indexes directories of vCards (*.vcd), allowing you to compose mail to the contacts inside.
> >
> > However, I'm not sure how to create the configuration for selecting multiple directories to monitor (currently it is just hard-coded for testing).
> > Also, folders aren't recursively monitored--I'm not sure of the best way to do that, nor am I sure if it is necessary.
> >
> > Could anyone help finish what I started?
> >
> > The plugin depends on the vobject module, which hopefully isn't an issue.
> >
> > Thanks!
> > -David
> 
> Hi David,
> 
> I forgot this email, I was not at home at that time.
> 
> What you propose suggests a new mechanism -- modifying how we
> understand files. Kupfer could have one central file indexer, and
> plugins would modify it by supplying parts to understand vCard files
> as contents for example. It's a pretty big/new idea but it would be
> interesting to explore.
> 
> Right now, Karol has some patches to supply plugins with the
> possibility of a custom settings dialog -- so with that, this plugin
> could have its own directory configuration.
> 
> Ulrik

Ulrik,

That makes a lot of sense.  For the sake of speed, perhaps plugins could register file extensions that they can processes; that way non-registered extensions can be trivially skipped over, keeping the scanner quick.  You lose the ability to handle misnamed or extensionless files, although I suppose a plugin that wants to check all files can somehow register itself for all files.  vCard handling could certainly be done trivially with this sort of system.
Any ideas for other handlers?
Some ideas (of varying worth, in no particular order):
 - ical (limited to future events to prevent an overabundance of items...but what would we do with the data?)
 - Music (can process tags such that artists/albums/titles are searchable)
 - Video (they less frequently have tags, but there's no reason we couldn't check for them)
 - OpenOffice documents, PDF documents, etc (extract document titles)
 - Plaintext files (search for words in the first line of files; shell scripts are excluded; maybe even hashtags in small files?)
 - HTML document (<title> tags, metadata tags)
 - .desktop files (sometimes they are in odd places, like the desktop)
 - Archive files (actually index the files inside, instead of just providing the ability to enter them manually via the latest plugin)
 - Code (searchable class names!  okay, now I'm just being silly.)

Care would have to be taken not to bog down the system...

-David