If remove_bad_urls is false, the database can contain
invalid URLs.
In CVS at 2003-10-14 on Mandrake Linux, htsearch
(correctly) suppresses matches for the names of these
URLs, but (incorrectly) counts them in the total number
of matches. This results in, for example, a page
saying "1-2 of 2 matches" but containing only one match.
For example, search for "technical" in the test suite
with remove_bad_urls=false.
Logged In: YES
user_id=624975
I found this bug when converting code to libhtdig.. Its an easy fix.
I'll commit the fix and have you test it.
Logged In: YES
user_id=624975
Lachlan:
Please try this patch in parser.cc
863,864d863
< DocumentRef *ref = NULL;
<
888,891d886
<
< ref = collection->getDocumentRef(dm->GetId());
< if(ref->DocState() == Reference_normal)
< {
896d890
< }
Thanks.
Logged In: YES
user_id=663373
I've applied, tested and committed the patch. Please close
the bug report once you have confirmed the patch was applied
correctly. Thanks :)
Logged In: YES
user_id=624975
Looks great. Thanks. Neal