* XmlNodeReaderImpl.cs : in GetNamespacesInScope(), avoid NREs under
"no current node" situation. Also, look for attributes on iterated
nodes, not current ones.
svn path=/trunk/mcs/; revision=160145
* XmlNodeReaderImpl.cs : when the reader is at attribute value,
MoveToNextAttribute() incorrectly moved to the first attribute.
Fixed bug #550379.
* XmlNodeReaderTests.cs : added test for bug #550379.
svn path=/trunk/mcs/; revision=145910
* *.cs: The tuner takes care of the visibility of this for us, and it
is in fact needed. (Readd ReadAttributeValue to 2.1)
svn path=/trunk/mcs/; revision=127366
* XmlReader.cs XmlNodeReader2.cs XmlTextReader.cs XmlTextReader2.cs
XmlNodeReaderImpl.cs EntityResolvingXmlReader.cs :
The previous attempt was wrong. Those internal-virtualized ones
could be used in System.Xml.Core classes and outside the assembly
any derived XmlReaders would have thrown throw Exception.
Now I marked them as nonexistent-in-2.1 or privatized.
* XmlWriter.cs : actually only this class is affected (if there
were DOM or XSD they would have been affected as well).
* SubtreeXmlReader.cs XmlFilterReader.cs : further dependent changes.
svn path=/trunk/mcs/; revision=83706
* XmlReader.cs : HasValue, ReadAttributeValue() and ResolveEntity()
do not exist in 2.1 profile. Hiding abstract stuff seems to cause
some issues at run time, so they are ifdef-ed out from 2.1.
* XmlNodeReader2.cs DTDValidatingReader2.cs XmlTextReader.cs
XmlTextReader2.cs XmlNodeReaderImpl.cs DTDValidatingReader.cs
EntityResolvingXmlReader.cs : hence the change above.
* SubtreeXmlReader.cs XmlFilterReader.cs : dependent changes due to
XmlReader.cs changes for 2.1 profile.
svn path=/trunk/mcs/; revision=83702
* XmlNodeReaderImpl.cs : Reimplemented Read(). Points are:
- reduced call to NextSibling which is very unlikely to be inlined
- reorganized conditioning to avoid extra check
- added description on each returning branch.
- And then, removed unused code.
It is 1.5x faster than before.
svn path=/trunk/mcs/; revision=57062
* XmlNodeReaderImpl.cs : do not compute ownerLinkedNode, just store
it in Read(). Simply use state field in EOF. They result in 20%
perf. boost. Expanding EOF also improved performance by 2-3%
(mono --optimize=inline might work in the later stage).
svn path=/trunk/mcs/; revision=56999
* XmlNodeReaderImpl.cs : when the source node was not a document and
it is an empty element, Read() was incorrectly moving to the parent
of the source node. Patch by Csaba Halasz. Fixed bug #76260.
* XmlNodeReaderTests.cs : added test for bug #76260.
svn path=/trunk/mcs/; revision=50962
* XmlReader.cs : ReadElementContentAsBinHex() was calling ..AsBase64().
* XmlValidatingReader.cs : In set_EntityHandling(), set that property
of DTDValidatingReader.
* XmlNodeReader2.cs,
XmlNodeReaderImpl.cs : added simplified implementation that splits
entity handling and anything else.
* XQueryConvert.cs : implemented ShouldCheckValueFacets().
Use Convert class instead of simple casts (check range).
* XmlReaderSettings.cs : added SetSchemas() for XPathNavigator.
svn path=/trunk/mcs/; revision=38088