||2002-12-22 17:04:52 GMT
||2004-05-16 18:49:11 GMT
(Projects of this owner)
REXML is an XML 1.0 compliant, reasonably fast, non-validating XML parser that supports multiple encodings. It has an API that is designed to be intuitive, straightforward, and terse. REXML includes a tree model parser, a SAX2 streaming parser, and a pull parser. It also includes a full XPath implementation. All of REXML's parsers pass 100% of the Oasis XML non-validating tests.
REXML is distributed with Ruby as of Ruby 1.8.
3.0.3 -> 3.0.8
* Fixes to transitive pretty-printing
* Attributes can now be non-Strings; they'll be to_s()'ed when set.
* Text now works on frozen strings
* Fixed a bug where, in some cases, the parser would fail to read enough characters from the input stream, and would consequently fail.
* In reporting a parse exception on an input stream, in some cases the code that calculated the position in the stream of the error would raise an IO error, which wasn't caught.
* Fixed Text so that using Elements outside of a Document wouldn't cause errors.
* US-ASCII encoding was incorrectly outputing 8-bit ASCII. (fix from Stefan Scholl)
* Mikko Tiihonen contributed encoding support for CP-1252 and ISO-8859-15
* Fixed XPath bug where any string equaled zero (ruby converts all non-numeric strings to 0, whereas XPath says they're NaN)
* Added a test for descendant-or-self ordering
* Added a test for deep cloning
* Added a getter for the source in BaseParser
* Replaced #inspect in both Element and Attribute. This should make stack traces nicer.
* Fixed a bug in ParseException. Line numbers in Ruby IO are a real pain in the ass!
* Functions::number is more correct
* Added Kou's backward-compatability fix for DocTypes using a Source as an argument to the constructor. This constructor has been deprecated; there isn't any reason
for it to exist, except for backwards compatability. I will probably remove it in the future.
* Fixed indentation with CDATA bug, #23