On 8 Aug 2000, at 7:12, Bill Moseley wrote:
> At 06:54 AM 08/08/00 -0700, Chris Humphries wrote:
> >> One more
> >> question, how we can restrict swishe to index only the meta tags in a
> >It would not be too difficult to write a small filter function to process
> >HTML documents so that only text from their meta tags was passed on as
> >document content (I have my own filter functions to process documents that
> >could lend themselves to this task).
> Just a thought:
> Another option that may or may not be faster than writing a filter would be
> to modify index.c. I wanted to be able to only index META tags that are
> defined in METANAMES so I added the REQMETANAME setting to config.h. It
> only took a few minutes to make the change in the source.
> As I remember (perhaps incorrectly), the index stores words in categories
> by number where the meta names are numbered 1,2,3... and everything else
> get is zero. (is that correct Jose?) Sorry, I don't have time to look at
> the source right now.
MetaNames are numbered 2,3,4... and everything else is 1. So
discarding non metanames is easy.
> So it might be easy to ignore non metanames in index.c.
Yes, it should be easy. Now, swish-e-2.0-beta4 is quite
stable (no more bugs have been reported). We can include this
feature in the next release as an option in the config file.
Received on Wed Aug 9 05:14:02 2000