This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
The search index does not appear to be being rebuilt - I believe it should be done once/day ? Currently, after running a search, the following appears at the bottom of the search results page : The search index is about 26 days old.
> Just want to get an idea if this is correct. The search engine hasn't > rebuilt the index for the past 26 days? Actually, it looks like it's been longer -- I think the last time it was rebuilt was : -rw-r--r-- 1 root root 25304905 Sep 22 17:30 swish.idx ... I'll add the appropriate crontab entry and see if that resolves this tomorrow. thanks -- Ed --
*** Bug 8414 has been marked as a duplicate of this bug. ***
The crontab entry didn't resolve it. We're working on it.
For CollabNet ppl reference, the issuezilla issue is http://projects.collab.net/issues/show_bug.cgi?id=1101
Edk's looking into it. He hopes to be able to look into it over the weekend.
--- Additional comments from Keri : > 1) Site Indexing -- Simply put, our site is too large to be indexed > using the indexer that we currently use (Swish-e). 3 developers have > been working pretty furiously behind the scenes to > a) come up with a mechanism to get Swish-e to work -- I think all of > those lines have been pretty much exhausted. > b) find a new indexer and set it up. Swish++ is available and is the > current line of investigation. If we're looking at replacing the search utility, we should probably nail down the features/functionality that we would like and need. Here's a start - - Must be able to selectively include or exclude specific sets of "stuff" to search. eg - I would like to have check boxes such that I can search *only* the nbdev mailing list archives. Or nbdev + nbusers, but not the www site. Or only www but no mailing lists. Etc. - should be able to not index useless stuff, like cvs msg archives. - would be nice to have a configurable .conf file or something that we have access to, and can tweak not only things like look and feel, but config options of the search utility itself (number of hits returned/page, ... etc)
Status as of today: Michael Stack is using Swish-e to index the site. He is indexing the site in small slivers and then merging the results. He is pretty confident that this will work and that the results will be posted late tonight or early tomorrow. This is based on an at least 15 min. conversation with him, following up through email and on IRC. It's later in the day and he still feels confident.
The index has been rebuilt and is live on Netbeans. Yay!!!
Indexes are being rebuilt, and working OK. However the returned hits include "project/www/" in the path, so for example following a hit I end up at http://www.netbeans.org/project/www/articles/31releasenotes.html instead of http://www.netbeans.org/articles/31releasenotes.html The first does actually work, but probably should not be used/visible to the user.
This is due to the way we deal with vhosts in the current instantiation. The two ways to reference URLS on Netbeans is known and will be worked on in future releases, but I'd rather not have to track it on a week to week basis at this time.
Fair enough - though since this is new behaviour, and a similar bug was fixed at the time of the move to this version of Tigris, I thought this might have been the same problem, reintroduced.
Not sure why this is market resolved "later". The original report of the srch index not being rebuilt was fixed, my suggestions for additonal search features are now included in seperate issues (#9721- #9724) ... Marking closed.
hmm, seems it ignored my closed/resolved, it is still closed/later. trying to close/resolve.
fixed
We recently moved out from Collabnet's infrastructure