Using Namazu 2.0.13 on Solaris. I built an index with:
mknmz --mhonarc -f /archives/cgi-bin/mknmzrc -T /archives/cgi-bin/template/ -O
/archives/html/ /archives/html/
However, single-word searches on this index often fail. If I restrict
to a subset /archives/html/ I get much better results (e.g. all
mharc-created mailing list indexes seem to be OK).
To get useful results, I have to use regexp-search. It appears that the
search is not exhaustive, which of course makes it useless. Here's an
example:
$ namazu -d the /archives/html/
namazu(debug): NAMAZUNORC: ''
namazu(debug): load_rcfile: /archives//etc/namazu/namazurc loaded
namazu(debug): -n: 20
namazu(debug): -w: 0
namazu(debug): query: [the]
namazu(debug): Index name [0]: /archives/html/
namazu(debug): set_phrase_trick: the
namazu(debug): set_regex_trick: the
namazu(debug): query.tokennum: 1
namazu(debug): query.tab[0]: the
namazu(debug): size of /archives/html//NMZ.t: 107652
namazu(debug): before nmz_strlower: [the]
namazu(debug): after nmz_strlower: [the]
namazu(debug): do WORD search
namazu(debug): size of /archives/html//NMZ.ii: 1436740
namazu(debug): l:0: !
namazu(debug): r:359184:
namazu(debug): searching: s
namazu(debug): searching: les
namazu(debug): searching: dterm
namazu(debug): searching: /rpc,
namazu(debug): searching: _exit
namazu(debug): searching: ishes
namazu(debug): searching: n.com
namazu(debug): searching: e:fs>
namazu(debug): searching: ldirs
namazu(debug): searching: 75
namazu(debug): searching:
namazu(debug): searching: done;
namazu(debug): searching: ~~~~~
namazu(debug): searching: ~~~~~
namazu(debug): searching: only;
namazu(debug): searching: first
namazu(debug): searching: ~~~~~
namazu(debug): searching: splay
namazu(debug): searching: y
Results:
References: [ the: 0 ]
No document matching your query.
Unfortunately I cannot give you the index files or access to them, but I
am willing to do whatever debugging is necessary. Any ideas?
thanks,
john