Implement all necessary features such as href() and getChildElement()
ourselves instead.
* Add missing classes to class.properties.
* Display actual line numbers in function source macro.
- TokenStream is no longer public, implement Parser.parseTokens()
- Parse public nested methods in JS constructors
- adapt to slightly changed sequence of returned tokens.
* Do not implement IPathElement in Server, Application and HelmaDoc classes.
Implement the necessary methods in the manage application instead.
* Add DocResourceElement.getStartLine() method to get element's position
within the containing resource
* Rename ScriptingEngine.getIntrospector() to getDoc() and declare DocApplication
as return value.
* Make DocFunction not a RuntimeException, and throw it only where needed.
* Add TokenStreamInjector class to rhino to make TokenStream publicly available
* Add StringUtils.splitLines()
* bugfix in getField(): don't try to convert the string value into a date, since this possibly does the conversion even if it's not intended (it can happen that an id is converted into a date). therefor re-introduced the method getDateField()
from http://dev.helma.org/Wiki/Rhino+Bugs+%26+Patches/:
- Rethrown exceptions rendered as [object Error]
- Provide protected callback methods in Rhino debugger SwingGui
- Bug in tail call optimization that cripples Rhino debugger
* Add Logger.Entry class to avoid unnecessary String concatenation.
* Unify Logger classes code a little bit, making better use of inheritance.
* Run logger thread 3 times per second instead of 4 times.
* Add Javadoc comments and tags.
* added (most of) missing analyzers in Search.getAnalyzer()
* createIndex doesn't support a third argument for force-create an index anymore. to clear an index use its create() method instead
* Index now uses an IndexModifier instead of an IndexWriter in most cases
* to create a new Searcher instance now call index.getSearcher() or pass the index to search in as argument to the Searcher constructor
* checkWriteLock() is gone, instead getWriter() and getModifier() try for 2 seconds to return a writer/modifier. this should avoid concurrency problems
* Index.addIndexes() now takes a number of index directory arguments
* Index.create() now deletes all files in the directory (before it didn't)
* added checks for existing writer/modifier/reader before calling close()
* Index.getFieldNames() now returns a javascript array instead of a java object one
* Index.isLocked() doesn't create a reader anymore (before it did, but didn't use it)
* removed Index.getTerms() - if someone is missing that, shout
* Index.addDocument() now expects a single Document instance, for batch adding use the new method addDocuments()
* same with Index.removeDocument()
* added new method Index.updateDocument() that does removal and adding in one step (so the index is locked just once instead of two times)
* the wrapped lucene Query instance is now a public field - this saves from duplicating the getter method in every extending prototype
* Document.addField() now uses new DateTools class instead of the deprecated DateField utility methods
* removed Document.getDateField() since Document.getField() now handles dates as well
* added JSDoc compatible documentation
* Remove old Rhino 1.5 classes, make Helma debugger subclass Rhino 1.6 classes
(not fully functional yet, needs patches to Rhino debugger).
* Make helma.doc.DocFunction compile with Rhino 1.6 by introducing intermediary
DockenStream class that proxies TokenStream using java reflection.