[GRLUG] gnome tracker issue

[Eric Beversluis]

On Thu, 2012-10-04 at 05:31 -0400, Adam Tauno Williams wrote:
> On Wed, 2012-10-03 at 20:31 -0400, Eric Beversluis wrote:
> > On Wed, 2012-10-03 at 18:14 -0400, Adam Tauno Williams wrote:
> > > On Wed, 2012-10-03 at 18:03 -0400, Eric Beversluis wrote:
> > > > It seems from some stuff on line that in fact tracker does not
> > > > index .php files.
> > > They at least make it into the index.  It indexes other types of code
> > > files.
> > > awilliam at linux-nysu:~> tracker-search jsonRPCClient
> > > Results:
> > >   file:///home/awilliam/Documents/Development/PHP/jsonRPCClient.php
> > >   file:///home/awilliam/Documents/Organizations/MI/OGo/CRM%20&%
> > > 20Hedera/TireSale/tiresale.py
> > > file:///home/awilliam/Documents/Works/Coils.MD/snurtle-code/build/lib/snurtle/jsonrpc.py file:///home/awilliam/Documents/Works/Coils.MD/snurtle-code/build/lib/snurtle/rpcclient.py
> > > ...
> 
> The root the of the issue here is that there is no extractor for PHP
> file types;  in Tracker an extractor exists for each file types and
> extracts the information from the file [ODT, DOCX, PDF, TXT, etc...] to
> be added to the index.  Extractors correspond to MIME types and the MIME
> type of a PHP file is, for some queer historical reason,
> "application/x-php".   Whereas Python files have a MIME type of
> "text/x-python" so if there isn't an extractor for "text/x-python" it
> falls back to extracting with "text/*".  Most code files are treated as
> text files: "text/x-c", "text/x-c++", "text/x-java", "text/js", etc...
> But PHP is "application/*" for which there is, obviously, not default
> extractor.
> 
> PHP developers, typically, being lower-case-d developers probably aren't
> going to write an extractor [which would be done in C].
> 
> You can try telling Tracker's text extractor that it can extract PHP
> files by editing /usr/share/tracker/extract-rules/90-text-generic.rule
> and changing the Mimetypes directive from "text/*" to
> "text/*;application.php;".  And then restarting tracker - logging in and
> out would probably do that.  Then just touch the PHP files, when the
> tracker miner becomes idle again [tracker-control] check one of the PHP
> files using tracker-info.
> 
> 
Thanks, Adam.
I'm giving that a try.
Should "text/*;application.php;" be "text/*;application/x-php"?
 Do I have to 'touch' all the .php files to get them to be indexed? Can
I do that with a wildcard ("touch *php")?

Missed you and Dirk at OLF.

EB