faq

A: Zilverline is free of charge for personal and R&D use. For commercial use you will have to pay or contribute. See Getting Zilverline.

Q: Zilverline asks me to login, what is the username/password?

A: admin/password for administration rights, indexer/password for indexing rights.

Q: I want to change these passwords, where do I do that?

A: You can change passwords in acegi-security.xml in zilverline/WEB-INF. Look for a line admin=password,… admin is the username, change 'password' into your stronger password. Same goes for indexer.

Q: I used to configure Zilverline through configuration files, has anything changed?

A: Since version 1.2.0 Zilverline is fully webbased. Only the theme and scheduling options are still in web-servlet.xml and schedule-context.xml respectively.

Q. Where is the configuration of the collections and all preferences stored?

A: The directory where the configuration files (collectionManager.xml and searchService.xml) are stored is by default your home directory (user.home), you can override this in WEB-INF/applicationContext.xml and WEB-INF/classes/dao.properties.

Q: Every night at 3:14 my machine starts running like a maniac?

A: Indexes can be created automatically by specifying a schedule in schedule-context.xml. The scheduling process starts (incremental) indexing every night at 03:14, which is set by default. Just uncomment the line and restart Zilverline to get rid of it. See scheduling.

Q: You say IMAP support is in alpha state. What does not work yet?

A: You can get most content from an IMAP mailserver, but the Handlers do not work yet. Also the results are shown as imap URLs, but browsers do not support that anymore.

Q: I'm having trouble with searching for non-western characters. When I submit 'á' I see 'á' returned. Is this a bug?

A: This is not a Zilverline issue. You are probably using Tomcat, and you have to instruct it to properly handle GETs with UTF-8 encoded parameters. See http://jakarta.apache.org/tomcat/tomcat-5.0-doc/config/http.html and have a look at the URIEncoding parameter. It basically means that you have to edit the server.xml:

<Connector port="8080" maxThreads="150" minSpareThreads="25" 
maxSpareThreads="75" enableLookups="false" redirectPort="8443"
acceptCount="100" debug="0" connectionTimeout="20000" disableUploadTimeout="true" URIEncoding="UTF-8" />

Q: Can I automatically update my collections?

A: Yes you can, either by scheduling or by using a SOAP webservice.

Q: What are the supported search fields (ie: text:, type:, title:)?

A: I'm using a few, you can actually look at them using luke, they are: summary, zipPath, type, title, zipName, hash, modified, contents, size, collection, path and name.

Q: How do I specify a more advanced query?

A: java "method invocation" rmi^6 +type:PDF +modified:[20020101 TO 20040701]
is an example that searches only PDF documents from Jan 1, 2002 to July 1, 2004, that hava java or 'method invocation' or rmi in them, where rmi scores 6 times as high. See Lucene Documentation for more information.

Q: In the collections view the checkbox is grey, I can't check it, and therefor can't create the index.

A: The checkbox is grey since the collection does not exist on disk. Check your collection definition and fix it. You can check the log file (as specified in log4j.properties) to see what's going on.

Q: How do I specify UNC style paths for my collections?

A: E.g. fill in \\fileserver\d$\books\example as the Content Location of yur collection.

Q: I use a Mozilla based browser, and the search result (starting with file://) from local disk does not open.

A: This is a security feature that prevents you from opening a local file from a remote page. Mozilla help explains what to do: http://www.mozilla.org/quality/networking/testing/filetests.html (look for "file:// links in HTTP document" at the bottom) For newer version see http://kb.mozillazine.org/Links_to_local_pages_don't_work#Firefox_1.5.2C_Mozilla_1.8.2C_and_newer

Q: How do I know what version of Zilverline I'm running?.

A: (since 1.0-rc4) Hover over the footer of a page (the part that says: Zilverline Search Engine, Copyright (c) 2004 Michael Franken).

Q: Something goes wrong, what can I do to find out what's going on?

A: To see what happens, or in case anything goes wrong, have a look at the log file (which can be configured through the log4j.properties). You can raise the log level to DEBUG. This will print very verbose information. Reload your webapp, or restart the servlet engine. Try what you were doing again, and have a look at the logfile again. If you still have no clue, send me an email with logfile, indexService.xml and collectionManager.xml attached.

Q: How do I raise the loglevel?

A: The loglevel can be configured through the log4j.properties. This file is localted in <webapps/zilverline>/WEB-INF/classes. You can raise the log level to DEBUG. This will print very verbose information. Reload your webapp, or restart the servlet engine. The logfile can be set, by default is it prints to stdout (so in Tomcat it will end up in logs/stdout.log) or else where your JVM started (C:\Windows\system32\zilverline.log ?).

Q: What platforms have been tested?

A: I've tested with Tomcat 4.1.x, 5.0.25, 5.5.16, Resin 3.0.7 and BEA Weblogic Server 8.1 using JDK 1.4 on Windows XP/2003 and Linux.

Q: What is a MD5 hash?

A: A MD5 hash is an unique code generated for a file, it is much like a CRC code, but believed to be more unique. It contains a 128 bit code. The MD5 hash is stored in the index.

Q: Can I index CHM files on Linux?

A: On Linux you can get chm support from here. Thanks to Nat Peterson for pointing to it.