Sophie: xpp2-doc-0:2.1.10-6.1.5mdv2008.1 x86

xpp2-doc-2.1.10-6.1.5mdv2008.1.x86_64.rpm

<HTML>
<HEAD>
<TITLE>Pull Parser 2 Changes
</TITLE>
   <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
   <meta name="Author" content="Aleksander Slominski [aslom@extreme.indiana.edu]">
</HEAD>
<BODY BGCOLOR="white">

<H1>Changes to Pull Parser 2 (XPP2)</H2><P>

<h3><a name="XPP_2_1_11"></a>2004- (XPP_2_1_11)</h2>

<ul>
<li>final maintenance release?</ul>

<P>

<h3><a name="XPP_2_1_10"></a>2003-11-15 (XPP_2_1_10)</h2>

<ul>
<li>[2003/9/4] fixed CDATA bug reported by Timo Thomas: Tokenizer class was
making assumption that pc[] is always equal in size to buf[] but this did not
hold in case when appending chunks of CDATA content (and not one character by
one character ...) refactored to encapsulate all System.arrayCopy calls with
extra checks to deal with this situation<li>added check to addDeclaredNamespaces to prevent adding of conflicting
namespace declarations for the same prefix in the XML node
<li>added equals() and hashCode() to implementation of Tag/StartTag/Node
and added unit tests for it
<li>fixed bug in CDATA handling in Tokenizer reported by Nikhil C. Khedkar and
added Junit test for it
</ul>

<h3>2003-03-07 (XPP_2_1_9)</h2>

<ul>
<li>added standard JAR output that has no JAXP11 or SAX2 driver as requested by Ben Walding
who wanted to be able to use JAXP11 parser and XPP2 and not worry about conflicts in CLASSPATH ..
<li>fixed bug in XmlNode.removeAttributeBy* that was showing if the attribute
was first removed and then added - slot at position <b>n</b> in attribute array
was incorrectly pointing to the same location as slot <b>n-1</b> ...
<li>changed readName to be more XML 1.0 and 1.1 compatible to allow full range of possible name characters
and not rely on Java Character.isLower/uppercase ...
<li><b>NOTE:</b> <em style="font-style: normal; font-weight: 700">backward incompatible change</em>:
fixed typo in method XmlStartTag.removeAtttributes -> removeAttributes name reported by Yogesh
<li>added to XmlStartTag two new methods: removeAttributeByName(String uri,
String localName) and removeAttributeByRawName(String rawName)<li>compiled and tested against latest Xerces 2 from CVS (bug with duplicated xmlns seems to be fixed in X2 CVS)
<li>improved X2 cumulative reader to correctly support un-shrinkable buffer
</ul>

<h3>2002-04-03 (XPP_2_1_8)</h2>

<ul>
<li>fixed a bug in tokenizer: it was trying always to fill buffer even though one byte
was still available - now it will use buffer to the end and then read new chunk from Reader
<li>updated XPP2 implementation to compile with patched latest Xerces2.0.1<li>
added build and run scripts (finally ...)</ul>

<h3>2001-12-10 (XPP_2_1_7)</h2>

<ul>
<li>fixed SAX2 driver to correctly report prefixes and better handle
invalid input for parsing
<li>improved Xerces 2 driver to use Augmentations (working with BETA3 of Xerces2)
</ul>

<h3>2001-11-21 (XPP_2_1_6_ALPHA)</h2>

<ul>
<li>fixed bug in default implementation of PullParser that prevented
it from reading list of declared namespaces when END_TAG is current event
<li>added SAX2 driver that translates pull events into SAX2 push callbacks
<li>added JAXP11 driver (based on code from Xerces 2 except for minimal changes)
</ul>

<h3>2001-11-20 (XPP_2_1_5_ALPHA)</h2>

<ul>
<li> split XmlPullParser interface into proper XmlPullParser
and separate "feature" interfaces
XmlPullParserBufferControl and XmlPullParserEventPosition
that clearly mark optional interfaces.
<br><b>NOTE:</b> for XmlPullParser pp
use <code>if(pp instanceof XmlPullParserBufferControl)</code>
to check if an optional interface is implemented by pull parser
</ul>

<h3>2001-10-18 (VER_2_1_4_ALPHA)</h2>

<ul>
<li>X2 driver is no integrated with latest Xerces 2 XNI
and the problem with EOL normalization is gone however there is
another problem related to attribute value normalization - reported
to Xerces 2 mailing list.
<li>added more tests and removed some namespace bugs
</ul>

<h3>2001-10-18 (VER_2_1_3_ALPHA)</h2>

<ul>
<li>X2 driver is becoming quite stable and passes almost all tests
except two reported bugs in Xerces 2 and lack of support for
low level parsing (control of input buffer and event position):
<ol>
<li> [major] handling of end-of-line normalization is incosistent with XML
spec ... (commented out checks in testNormalizeLine() from ParserTest for
now)
<li> [minor] apparently Xerces allows duplicate attribute with namespaces to
be declared (commented one line checking it in testAttribUniq from
ParserTest for now)
</ol>
<li>it is also very easy to switch between X2 driver and default XPP2 driver
implementation  or even control it by system property (similarly to JAXP).
when both drivers are on CLASSPATH normally first on CLASSPATH will be
used but the selection can be forced  with system property by writing:
<code>XmlPullParserFactory.newInstance(System.getProperty(XmlPullParserFactory.DEFAULT_PROPERTY_NAME));</code>

</ul>


<h3>2001-10-10 (VER_2_1_2_ALPHA)</h2>

<ul>
<li>a first milestone towards support of other parser engines
in XPP2 and SAX2 driver. initial X2 driver can run simple count
sample but does not pass internal XPP2 unit tests (yet!)
</ul>

<h3>2001-08-20 (XPP_VER_2_0_2 is the same as VER_2_0_1_BETA8)</h2>

<p><strong>Stable 2.0.2 version released and frozen!</strong>
Summary of changes since initial 2.x version:<ul>
<li>reduced memory footprint of node tree by 50% (both Node and PullNode),
<li>increased performance tree node build by 20%
<li>improved and optimized output (Formatter and Recorder)
<li>much more of JUnit tests
<li>added PalmOS Midlet sample
<li>small changes and improvement in API
<li>fixed few bugs ...
</ul>

<h3>2001-08-20 (VER_2_0_1_BETA8)</h2>

<ul>
<li>fixed bug that prevented to add first attribute
to StartTag (ensureCapacity called with 2*size where size == 0 ... )
reported by Dennis Sosnoski.
bug was introduced when fine tuning memory footprint in BETA6...
<li>XmlPullNode.getPullParser() will now
correctly parse unfinished children before returning parser,
add JUnit tests...
<li>changed XmlPullParser.getCurrentEvent() to getEventType()
<li>switch to use in build.xml tar/gzip instead of zip
(much smaller output archive!)
</ul>

<h3>2001-08-19 (VER_2_0_1_BETA7)</h2>

<ul>
<li>changed Node implementation to use
smaller Vector of initial size 4 and increment of 4
<li>pull parser is no longer reporting "" content
</ul>

<h3>2001-08-18 (VER2_0_1_BETA6)</h2>

<ul>
<li>reduced by half tree node memory
footprint by avoiding placeholder attributes
(XmlNode and XmlPullNode)
<li>rearranged XPP2 implementation into separate packages:
node, tag, pullnode, tokenizer, pullparser,
format
<li>fixed duplicate calls to readNodeWithoutChildren(...)
in PullNode (when constructing and just after that)
<li>added XmlStartTag.isAttributeNamespaceDeclaration()
to allow to avoid duplicating namespace declaration attributes
in Recorder/Formatter output when pull parser is returning
all attributes setNamespaceAttributesReporting(true)
</ul>

<h3>2001-08-16 (VER_2_0_1_BETA5)</h2>

<ul>
<li>Recorder will no longer print default namespace declaration
on top level element as it is default in XML anyway
(so no xmlns="" output)
<li>empty nodes will be output by Recorder as empty elements (ex. &lt;empty/>)
<li>escaping \t and \n in attributes
to preserve attributes values against normalization
<li>during parsing when constructing Node and PullNode
empty strings ("") will not be added as children
(they are not needed and we can save some memory)
</ul>

<h3>2001-08-16 (VER_2_0_1_BETA4)</h2>

<ul>
<li>fixed bug reported by Dennis Sosnoski in implementation of Node
that incorrectly reported getChildAt(0) when removeChildrenAt(0) was called before
</ul>

<h3>2001-08-15 (VER_2_0_1_BETA3)</h2>

<ul>
<li>added Palm OS MIDP built target for generating PRC, updated J2ME doc
with sample screen dumps
</ul>

<h3>2001-08-14 (VER_2_0_1_BETA2)</h2>

<ul>
<li>fixed bug reported by Dennis Sosnoski that caused Formatter
to recursively escape XML
</ul>

<h3>2001-08-12 (VER_2_0_1_BETA)</h2>

<ul>
<li>worked out J2ME build quirks and got XppCountMidlet to run - finally...
<li>finished PullNode implementation
</ul>

<p>&nbsp;<p>[<a href="../README.html">Back To Pull Parser 2 Documentation</a>]<p>&nbsp;

<HR>
<address><a href="aslom@extreme.indiana.edu">Aleksander Slominski</a><address>
</address>
</BODY>
</HTML>