* XmlChar.cs : (IndexOfInvalidChar) fixed surrogate pairs handling.
They were always regarded as invalid.
* XmlTextWriter2.cs : seems like MS XmlTextWriter outputs invalid
characters as character references. It totally does not make sense
but corcompare depended on that behavior.
This "compatibility" fix causes 10% or more performance loss!!!
* XmlTextWriterTests.cs : added invalid character processing test.
svn path=/trunk/mcs/; revision=57145
* XmlChar.cs : added IndexOfNonWhitespace() and IndexOfInvalid() (not
in use yet).
* NewLineHandling.cs : prepared to be used internally in 1.x.
svn path=/trunk/mcs/; revision=56943
* XmlChar.cs, XmlConstructs.cs :
Switched again, for reducing per-process memory consumption.
* XmlInputStream.cs :
Just 64 bytes of buffer would be enough, for example like
<?xml version="1.0" encoding="iso-8859-15" standalone="yes" ?>
svn path=/trunk/mcs/; revision=35382
* ChangeLog : Fixed incorrect description that should be written for
DTDValidatingReader.cs but written for DTDReader.cs.
* DTDValidatingReader.cs : Fixes described in the last DTDReader fix.
* XmlChar.cs,
XmlConstructs.cs : Made XmlChar and XmlConstructs equivalent and
replaced XmlChar with XmlConstruts, renaming "XmlChar" to
"XmlCharCompact" and "XmlConstructs" to "XmlChar".
(XmlChar will be used for compact framework.)
* DTDReader.cs, XmlParserInput.cs, XmlTextReader.cs :
XmlChar related changes. Support for surrogate pair.
It cannot use TextReader's peek anymore (for surrogate pair).
Character range check for PI value, attribute value, text
* XmlTextWriter.cs : Change openElements from Stack to ArrayList.
(It calls ToArray() every time.)
svn path=/trunk/mcs/; revision=22741
* XmlChar.cs : now GetPredefinedEntity() returns int and -1 for not
predefined entity names.
* DTDObjectModel.cs : DTDxxxCollection now derives from
- DTDCollectionBase that derives from DictionaryBase.
- "Literal value" and "replacement text" are handled differently.
- Throw XmlException when external entity could not be resolved.
- Return ReplacementText for EntityValue.
* DTDReader.cs :
- Support IXmlLineInfo.
- Fixed IGNORE section handling to be more strict.
- Compute replacement text on reading entity declarations.
- Don't expand char references before getting literal entity value.
- Check not allowed parameter entity inside internal subset.
- Added ValidateWhitespaceNode() to check whitespace's Validity
Constraints on external elementdecl and standalone document.
- When invalid data against its datatype was found, don't throw
parse error.
- Other code refactory.
* XmlTextReader.cs :
- Check for invalid Text pattern "]]>" should not depend on line
info since it might not be supported.
- Check referenced character's validity (WFC).
svn path=/trunk/mcs/; revision=22730
* System.Xml.dll.source: add System.Xml/HighWaterStack.cs
* XmlChar.cs (IsWhitespace): Spare enumerator allocation.
* XmlTextWriter.cs (WriteWhitespace): Enumerators are very high in
fat, use low-fat loops instead. Even better, call the function
inside XmlChar for code reuse.
* HighWaterStack.cs (added): New class that acts like a Stack, but
helps with memory allocation.
* XmlNamespaceManager.cs, XmlParserContext.cs: Rewrote to use
HighWaterStack.
* XmlTextReader.cs: Remove orderedAttributesEnumerator variable,
use a variable that stores the position in the ArrayList.
svn path=/trunk/mcs/; revision=17385
* DTDObjectModel.cs : added XmlResolver related members. Added
invalid entity recursion logic. Added encodingdecl check on textdecl.
* DTDValidatingReader.cs : It now implements IHasXmlParserContext.
* XmlChar.cs : IsPubidChar() should not allow TAB(	).
* XmlDocumentType.cs : 1) internal CreateDocumentType() was now changed
to receive DTDObjectModel to support other than XmlTextReader.
2) Most of its public member is now based on DTDObjectModel.
* XmlDocument.cs : 1) Synchronous change with XmlDocumentType.
2) ReadNode() now considers XmlParserContext's DTDObjectModel for
other than XmlTextReader (such as XmlValidatingReader).
* XmlNode.cs : code cleanup only.
* XmlParserInput.cs : added HasPEBuffer, used to check illegal nesting.
* XmlTextReader.cs : 1) Illegal entity reference check logic was moved
from ReadContent() to SetEntityReferenceProperties(). 2) Indentation
change on ReadEntityReference(). 3) ReadAttribute() now checks
reference to external entity reference. 4) Added textdecl encoding
check. 5) DTDObjectModel fields are now set correctly. 6) added
PERef markup nest check. 7) If PEDecl was not found, it might be WFC
violation, not only be VC violation. 8) ReadEntityDecl() now receives
its declared entity itself, and this method checks IsInternalSubset.
* XmlValidatingReader.cs : 1) Added GetInternalPerserContext().
2) ValidationType.None should be the same as Auto, not DTD (in the
future it should keep xml schema's default values).
Pending Stuff in XmlTextReader which breaks some NUnit tests;
1) SetEntityReferenceProperies() has check for illegal references.
2) ReadAttribute(bool) has similar check for illegal references.
svn path=/trunk/mcs/; revision=17218
* XmlInputStream.cs :
Fixed Initialize() that might result in incorrect buffer.
CanRead and Position should consider buffer (especially when the
stream was EOF).
CanSeek should use stream's CanSeek.
* XmlChar.cs : added GetPredefinedEntity().
* DTDObjectModel.cs :
- Modified ComputeDefaultValue() to use it.
- Fixed NormalizedDefaultValue not to expect parsed value as string.
- ResolveExternalEntity now detects invalid standalone specification.
- DTDParameterEntityDeclaration.Value is now changed to property and
keep its resolved value of literal value of external value using
new Resolve (XmlResolver) method (called only by XmlTextReader).
* DTDValidatingReader.cs :
- It now handles entity-expanded Whitespace and SignificantWhitespace
correctly. Does not raise unexpected text not allowed error.
- Fixed MoveToAttribute()s to move internal reader correctly, and
should be able to move in other cases than on element node.
- On Read()ing document type and ResolveEntity(), it now sets
XmlResolver on text reader.
- On ID constraints check and type verification, it now uses
normalized attribute or its tokenized value(s).
- Fixed ValidateAttributes() to add default values when
ValidationType is Auto.
* XmlParserContext.cs : added internal Dtd setter.
* XmlParserInput.cs :
InsertParameterEntityBuffer() should add surrounding whitespaces.
Removed unused code block.
* XmlTextReader.cs :
- Modified ReadReference(), ReadAttribute() and Dereference() to use
XmlChar.GetPredefinedEntity().
- When returning entity reference, it checks extity declaration
existence in certain conditions.
- ReadAttribute() now skips general entity replacement on reading
entity value.
- ReadText() now handles whitespace texts as XmlNodeType.Whitespace.
- ReadXmlDeclaration() holds isStandalone, which might be used in
entity reference's well-formedness check.
- internal 'DTD' now uses XmlParserContext.Dtd.
- isInternalSubset was meaningless. Now uses input stack's Count.
- Added some required SkipWhitespace().
- Several parameter entity handling changes. Added GetPEValue(name).
TryExpandPERef() now uses ImportAsPERef(). ReadParameterEntityDecl()
now uses its Resolve() when it was external PE. Now it uses value
buffer instead of currentTag, to efficiently read (included) PE.
CompileDeclaration() now uses ImportAsPERef instead of ExpandPERef().
ImportAsPERef() simply inserts the value to currentInput.
ReadEntityDecl() also uses value buffer. Removed non-used methods.
- Default attributes are now stored in normalized form.
- Dereference() now considers non-expanding predefined entities.
* XmlValidatingReader.cs : now holds Schemas, and ValidationType.Auto
is more correctly supported.
* XmlAttributeCollection.cs :
When removing default attribute, it immediately inserts the attribute.
RemoveAll() should consider default (not-removable) attributes.
* XmlDocument.cs :
Should have its default XmlResolver. Added Internal Resolver.
Some Load() should use its XmlResolver.
ImportNode() should copy its children in node level, not value level.
It also have to consider not to copy default attribute on importing
XmlElement.
* XmlDocumentType.cs : Now uses document's XmlResolver to read DTD.
* XmlAttribute.cs,
XmlDocumentFragment.cs.
XmlElement.cs : use XmlResolver on InnerXml. Removed some MonoTODO.
* XmlNodeReader.cs : GetInternalParserConext() should provide its DTD.
ResolveEntity() uses XmlResolver on entityReader.
svn path=/trunk/mcs/; revision=17202
* DTDObjectModel.cs :
- Added validation error check (and AddError(), Errors).
- Fixed ComputeDefaultValue() to handle various references correctly.
- DTDEntityDeclaration.EntityValue became property, and added
LiteralEntityValue. The new one holds resolved value.
Added ResolveExternalEntity(). It now required root in .ctor().
* DTDValidatingReader.cs :
- Now it handles namespaced attributes (as input to xsd validator).
- Added XmlResolver property as usual XmlReader.
- Added currentEntityHandling field so that it can stand changing
XmlValidatingReader's EntityHandling dynamically.
- FilterNormalization() now requires name for getting datatypes and
can stand for non-current attribute normalization.
- Splitted ReadContent() from Read() that can be called recursively
when expanding entities.
- Now handles Entity not found error *after* resolution of entities,
as MS.NET does.
- Read()ing DTD checks its Errors and raises validation error events.
In some situations, DTD parsing may detect VC error, not WFC error.
It also strictly checks NData existence.
- Handling of entity-resolved text is a bit changed and Read()ing
element, endElement, cdata now changed to switch collecting text
and collected text.
- content type ANY should allow texts.
- Added enumerated attribute validity check.
- Added Name/Names creation rule check for ID/IDREF/IDREFS.
- Added entity existence check for ENTITY/ENTITIES attributes.
- Added NMTOKEN creation rule check for NMTOKEN/NMTOKENS.
- Fixed to remove extraneous #REQUIRED check.
- Contributing default attribute is now only applied to the case
ValidationType is DTD or None.
- ResolveEntity() now handles external entities.
Added Mono.Xml.IXmlParserContext interface support.
* XmlDocumentType.cs : Fixed to use BaseURI to build DTD model.
* XmlTextReader.cs :
- Fixed ReadAttributeValue() to reset returnEntityReference.
- Added bool MaybeTextDecl only for ResolveEntity().
- Fixed ReadWhitespace(). In fact its value is considered even if it
is in the end of the XML.
- Fixed ReadXmlDeclaration(). Now skips text declaration.
- CompileDTDSubset() now checks IGNORE/INCLUDE section end balances.
- ReadContentSpec() should set OrderType = OR for mixed model.
Fixed to set content element name correctly.
- ImportAsPERef() now skips Text declaration.
- Type of enumerated default attributes shold be string.
- Undeclared PE error is now handled as DTD's VC error, not
XmlTextReader's WFC error.
* XmlValidatingReader.cs : BaseURI should provide that of
original XmlReader's when read was not started yet.
Fixed XmlResolver to delegate to DTDValidatingReader.
* XmlAttributeCollection.cs : Acquiring attList declaration is
insufficient especially in case of lack of elementdecl.
* XmlChar.cs : (Maybe this is compact than XmlConstructs.)
added IsWhitespace(),IsNCNameChar(),IsName(),IsNCName(),IsNmToken(),
IsPubid(). Copies IsValidIANAEncoding() from XmlConstructs.
* XmlInputStream.cs : Removed "version" declaration check. It should
be done by XmlTextReader.
* XmlNodeReader.cs : added internal GetInternalParserContext() and now
it can be used in XmlValidatingReader.ResolveEntity().
ResolveEntity() sets XmlTextReader.MaybeTextDecl.
svn path=/trunk/mcs/; revision=17167