Log Message:
Update to 3.98.1.1

Upstream changes:
Version 3.98-1

  *      Compilation error with clang. Simple declaration of a routine.

Version 3.98-0

  *      Update for libxml2-2.9.1 and reading from a connection for xmlEventParse().

  *      xmlIncludes() is a hierarchical version of getXIncludes()

  *      Modifications to xmlSource(), e.g. verbose = TRUE as default.

Version 3.97-0

  *      Fix for xmlValue(node) = text.  Identified by Lawrence Edwards.
         Uses xmlNodeSetContent() now and leaves freeing the original content to \ 
that routine.

  *      Updates for xmlSource()

Version 3.96-1

  *     readHTMLTable() ignores headers that are over 999 characters.

  *     Fix a problem in readHTMLTable() with some table headers not having
        the correct number of elements to match the columns.

Version 3.96-0

  *     Introduced readHTMLList(), getHTMLLinks(), getHTMLExternalFiles(), \ 
getXIncludes().

  *     When serializing XMLNode objects, i.e. R representations of nodes, \ 
ensure " and <, etc. in attributes
        are serialized correctly.

Version 3.95-1

  *     Allow htmlParse(), xmlParse(), etc. ?

Version 3.95-0

  *     Moved development version of the source code for the package to github -
        https://github.com/omegahat/XML.git

  *     Changes to the structure of the package to allow installation directly \ 
rather than
        via a one-step staging into the R package structure.

  *     Sample XML documents moved from data/ to exampleData, and examples updated.

  *     getDefaultNamespace() and matchNamespaces() use simplify = TRUE to call
        xmlNamespaceDefinitions() to get the namespaces as a character vector \ 
rather than
        list.

  *     Documentation updates

Version 3.94-0

  *     getNodeLocation() now reports the actual line number for text nodes \ 
rather than 0,
        using the sibling nodes' or parent node's line number.

  *     xpathApply() and related functions work with builtin type \ 
"functions",
        e.g. class.

  *     xpathApply() and related functions (getNodeSet, xpathSApply) allow
        the caller to specify multiple queries as a character vector
        and these are pasted together as compound location paths by
        separating them with a '|'.  This makes it easier for the
        caller to manage the different queries.

  *     assigning to a child of a node works, e.g. node[["abc"]] = \ 
text/node
        and node[[index]] = text/node.  We replace a matching name.  If the
        replacement value is text, we use the name to

  *     getChildrenStrings() is a function that implements the equivalent of
        xmlApply(node, xmlValue) but faster because we avoid the function call
        for each element.

  *     options parameter for xmlParse() and htmlParse() for controlling the parser.
        (Currently only used when encoding is explicitly specified.)

  *     encoding parameter for xmlParse() and xmlTreeParse() now works for XML \ 
documents,
        not just HTML documents.

  *     Update for readHTMLTable() method so that we look at just the final \ 
<tr> node
        in a <thead>.

Version 3.93-1

  *     Fixed bug in findXInclude() that sometimes got the wrong \ 
XMLXIncludeStartNode.
        Hence getNodeLocation() might report the wrong file, but correct line number!

  *     findXInclude() now has a recursive parameter that resolves the chain of \ 
XIncludes.
        This returns the full path to the file, relative to the base/top-level \ 
document,
        not just the parent document.

  *     Change to the default value of the error parameter in htmlParse() and \ 
htmlTreeParse()
        which will generate a structured R error if there is an IO error.
        The set of issues that will raise an error will be broadened in the future.

Version 3.93-0

  *    Enabled the fixing of <dummy> namespaces by finding the definition o
       for that prefix in the ancestor nodes.

Version 3.92-2

  *    Synchronized compilation flags for Windows with those on OSX & Linux.

Version 3.92-1

  *    Restore original error handler function for htmlParse() and htmlTreeParse()

  *    Fixed a reference counting problem caused by not adding a finalizer in the
       as() method for coercing an XMLInternalNode to an XMLInternalDocument.
       Example from Janko Thyson.

  *    Fixed up some partial argument names found by R CMD check!

Version 3.92-0

  *    Added --enable-xml-debug option for the configure script and this activates
       the debugging diagnostic reporting, mainly for the garbage collection and node
       reference counts.

  *    Work-around for HTML documents not being freed (but XML documents are!)

  *    Added an isHTML parameter for xmlTreeParse.

  *    Merge htmlTreeParse/htmlParse with xmlTreeParse.

  *    Implemented some diagnostic facilities to determine if an external pointer
       is in R's weak references list.   This needs support within R. (Ask for \ 
code if
       you want.)

Version 3.91-0

  *    Start of implementation to allow nested calls to newXMLNode() to use \ 
namespace prefixes
       defined in ancestor nodes.  Disabled at present.