Tobias Freitag wrote:
> 
> Alexander Barkov schrieb:
> >
> > Tobias Freitag wrote:
> > >
> > > Alexander Barkov schrieb:
> > > >
> > > > Tobias Freitag wrote:
> > > > >
> > > > > The problem is that when I search for any pattern, most results show
> > > > > directories not files. It seems that the indexer is parsing the
> > >
> > > > >  ...
> > > > > UrlPathWeight -1
> > > > > UrlFileWeight -1
> > > > >  ...
> > > >
> > > > This is because of -1 in UrlFileWeight and UrlPathWeight.
> > >
> > > I have set both values to 0 and to 1 but it doesnt change anything. Is
> > > this a common problem?
> >
> > Set it to 0 and reindex everything.
> 
> I did that already, but it didnt help.
> 
> I also cleared the whole database before indexing (using "echo YES |
> sbin/indexer -C ; sbin/indexer"). Maybe its a strange behavior under Red
> Hat 7?

The only thing I can guess is that those documents do have words you 
are trying to search.


> Is it possible to index with just \.html$ \.htm$ \.txt$ allowed? And why
> does the programm ignore the CheckOnly Entry?

Check that you have "regex" keyword in allow/disallow/checkonly
commands.
By default string wildcards with * and ? signs, not regular expressions
are used.


> And last but not least: Is there a difference between
> \.html$|\.htm$|\.txt$ and \.html$ \.htm$ \.txt$ ?


There is in some cases.

This:
   Disallow regex NoMatch \.html$|\.htm$|\.txt$
   Allow *

  means "index html, htm and txt extensions and skip others"


This:

   Disallow regex NoMatch \.html$  \.htm$  \.txt$
   Allow *

is just trunslated to:


   Disallow regex NoMatch \.html$
   Disallow regex NoMatch \.htm$
   Disallow regex NoMatch \.txt$
   Allow *


   Imagine that indexer decide what to do with for example a.txt. It
will find 
   the first command 
"Dislallow regex NoMatch \.html$". a.txt does not match 
   this regexp, so indexer will decide to disallow a.txt.  This
combination will 
   allow to index only .html files. Compare with the first one.
______________
If you want to unsubscribe send "unsubscribe udmsearch"
to [EMAIL PROTECTED]

Reply via email to