store reference to SearcherManager in IndexSearcher super-class fixes #1207

messaging addition

try to make freeing of IndexSearcher objects more robust

fixing clear button

hiding previous results after search click

fixing inconsistent behaviour of opengrok cookie

display last modified date in search results fixes #590

show time to get the results even when no match was found fixes #698

cleanup spellchecker code use Direct spellchecker to avoid another index creation closes #637

field type adjustments, ui fixes

Improve escaping of strings in webapp. Util.java: Improve formQuoteEscape() to also escape '&' characters, and clarify that it's for use in HTML attribute values. UtilTest.java: Add test for the new behaviour of formQuoteEscape(). menu.jspf: Use correct form of escaping for file type selection. search.jspf: Escape previously unescaped strings.

Added new field “type” to the web search form. The field is a single selection containing file types with user readable descriptions. It is mapped to the field “fileType” in Lucene.

Use description of sort order, not its name, in the search result. Changes "sorted by relevancy" to "sorted by relevance" "sorted by lastmodtime" to "sorted by last modified time" "sorted by fullpath" to "sorted by path"

HTML tags in parse error help text should not be escaped

Fix Checkstyle warnings (FileTabCharacterCheck)

Refactor webapp

Backed out changeset 813a4561226a Use a Filter instead!

RFE 18242: Add callbacks to allow people to filter access This is an initial version of the access filter. We might need to add checks in more places in the jsf files (they are really looking ugly.. we should clean them up!)

Bug #17936: ArrayIndexOutOfBoundsException when using n=50 as URL query parameter

Bug #17631: Cookie name "OpenGrok/project" is a reserved token Renamed cookie from OpenGrok/sorting to OpenGrokSorting.

Bug #17550: Malformed search for query with national chars

Bug 13184 - provide quick links to annotate,history and download from search results view

Bug #17127: OpenGrok leaving too many open files

Bug #17074: Home link no longer works Append "/" to the home link to account for getContextPath() returning an empty string if the webapp is deployed in the root directory.

Bug #16692: Index files left open after multiple project search

upgrade lucene to 3.0.2 fix regression on search for a definition which doesn't have a tag in index (e.g. classes just in jar, not in source)

Bug #15637: query constructed only for AND operators

Bug #15636: wildcard search for definitions does not work

Bug #14978 - upgrade lucene to 3.0.1

Bug #3435: C functions are not considered as functions Don't generate a direct link to the definition of a symbol unless there is only one definition of that symbol.

Bug #3900: Unexpected result when clicking a symbol definition

upgrade to lucene 3.0.0

tiny fix for null exception if project box contains rubbish

fix a regression with xref project generation, checkstyle fix, search link fix from search (s? comes from xrefs, search? comes from main search)

redesign project field, jquery used to autocomplete, fix findbugs warnings, make all pages http://validator.w3.org/ valid

increase the version number to 0.8 (finally ;) ) , reuseToken is inherited now (less code in jflex classes), small fixes in help and logic of showing the xrefs if file exists, it's xref gzipped exits and there is another indexed file with file.gz too (e.g. testdata directory)

Bug 341 - RFE: support searches across multiple projects some small UTF-8 fixes (select box still didn't pick them up in some test cases where fs was not UTF-8) validated default styles

small regression with 6th page of results after paging in lucene is used :)

lucene upgrade phase 2 , patch for 7594 - add support for sorting results by filepath by Christian Lohmaier

Bug #3561: "Did you mean:" should give a suggestion, or not be displayed.

Bug #2395 Searching fails for expressions starting by special characters

New batch of Findbugs fixes

Bug #2367: Colon in path property causes search.jsp to function incorrectly

Bug #334: double-colons in search pattern results in error Replace all occurences of : with \: in definition, symbol, path and history. If you would like to use "advanced" search (by specifying the fields yourself) you must use the "Full Search" box.

Use separate Lucene index and spelling suggestion database pr project

File links on the search page were percent encoding the segment separators ('/'), breaking the links. Created URIEncodePath that preserves the separators and percent encodes the segments. Fixed minor RFC 3986 deviations sections 2.1 (hex A-F should be uppercase) and 2.3 (unreserved chars should not be percent encoded) in URIEncode. Fixed places where the only the last segment was being percent encoded.

Fixed compile errors in search.jsp

Encode the filename in the links to allow it to contain #

Fixed compile error

Create webapp as part of the normal build step The webapp will also include the various configurations, so the user may select the look'n'feel without rebuilding the application. By default the webapp will use the look'n'feel (css-files and icons) located in the folder named default. You may alter the content in this directory, or specify the look'n'feel to use by using the -L option and create a configuration file to OpenGrok. ex: java -jar opengrok.jar -R configuration.xml -W configuration.xml \ -U localhost:2424 -L offwhite -n will read the current configuration, set the look'n'feel to "offwhite" and store the new configuration. The -U option I specified here was to update the webapplicaton that listens to configuration changes on localhost:2424

URL prefixes were always '/source?s' in the web application

Add support for leading wildcards Unless you specify a value, OpenGrok will use the same default value as Lucene (this is currently disabled). The current implementation will give you a performance penalty when you perform a search with leading wildcard.

Temporary fix for search results with only one hit If a search just return one hit, send the result page with only one hit instead of redirecting to the file unless we know the anchor in that page (Only true for TermQyery). If you run with projects enabled most queries will be BooleanQuery. This should be improved to determine the correct anchor and redirect to the correct page.

Update Lucene to 2.1.0 and a small configuration change * It is possible to specify the maximum number of words in a file to index by using: -b <number>. The default value is 60000

Store sorting preferences in a cookie

Added optional sort by last modified time

Fixed bug when searching with no project selected.

Removed multiple copies of webpages Some of the files stored in the conf/*/html directory did not contain any configuration specific text, so they could be moved to the web directory.

Prototype for "Project" support First of all: This is a prototype, so don't expect it to be finished ;-) With this version you should be able to specify various "projects" that you would like to group your software into, and limit your search within the selected projects. Did I mention that this is a prototype? ok.. The history cache has been moved into $DATA_ROOT/historycache, and if you're using Mercurial, you should probably pass the -H option to the indexer when you generate the index database...

Added a meta robots directive to the page headers

change foot.html to foot.jsp and use <context>/img/servedby.png for the image

Initial import (OpenGrok 0.3.1)