Colorer Library take5.alpha3

1 March 2003

BaseEditor class becomes more independent. It is based on ParserFactory class and allows now to internally create and choose requeried RegionMapper classes. Also, there is no needs now to load HRC FileType manually. If not yet loaded, any FileType class is loaded automatically, when passed to BaseEditor instance.
Java language interface is now avaiable. It uses Colorer through JNI calls. It defines mappings of BaseEditor and ParserFactory classes. Also simple Eclipse Platform plugin were written to demonstrate Colorer's abilities. It uses Eclipse's SWT API and can be used also as independent application.
Keywords sorting algorithm is replaced with simple quicksort method. This makes good performance increase when keywords in HRC files were not sorted.
Some changes in the internal handlers structure. LineSource class must pass now string pointer to parser, not DString object, as it was before. Also, now it defines formal requirements of the lifetime of the returned string. RegionHandler interface provides now references to currently parsed line of text. This makes easie installation and structure of external installed handlers.
LineSource class has now two new methods - startJob and endJob. They are used to notify object about start and end of line requests blocks.

HRC Language Reference is started.
Relax NG 1.0 schema is included.
XSLT 2.0 Draft schema is included (needs fixes).
weight attribute of firstline and filename RE matchers in prototype declaration has floating point type now. This makes type detection in some cases more flexible.
Automatic outlines generation option from xml schemas. All standard languages were regenerated to include common outline regions indention.
XHTML syntax uses Strict module by default and Transitional as alternative.
DocBook now uses latest correct XSD descriptions version 1.0b1.

HTML output mode uses font style information and can format sources now with bold, italic, underline properties.

clr: prefix allows to to use double quotes in file name.
Much faster performance (x2) was achieved int FAR's editor because of previously used multiple requests of each FAR's line. This makes great leak of cpu usage. Also, effective method of line text requesting is used in all possible actions.
Internal Library text lines requests has changed to make processing faster. Now, FAR plugin works faster because less lines data requests.
Structured presentation of created outliner. All elements are displayed in tree, also, for each element, prefix-char is shown. You can collapse or extend tree with Ctrl-Left/Right.