1 * XML Classes Status and Tasks
5 XML library is used by several areas of Mono such as ADO.NET and XML
6 Digital Signature (xmldsig). Here I write about System.Xml.dll and
7 related tools. This page won't include any classes which are in other
8 assemblies such as XmlDataDocument.
10 Note that current corlib has its own XML parser class (Mono.Xml.MiniParser).
12 Basically System.XML.dll feature is almost finished, so I write this
13 document mainly for bugs and improvement hints.
17 *** System.Xml namespace
20 **** Document Object Model (Core)
22 DOM implementation has finished and our DOM implementation scores better
23 than MS.NET as to the NIST DOM test results (it is ported by Mainsoft
24 hackers and in our unit tests).
28 Here XmlWriter almost equals to XmlTextWriter. If you want to see
29 another implementation, check XmlNodeWriter.cs and DTMXPathDocumentWriter.cs
30 in System.XML sources.
32 XmlTextWriter is completed, though it looks a bit slower than MS.NET (I
37 XmlUrlResolver is implemented.
39 XmlSecureResolver, which is introduced in MS .NET Framework 1.1 is basically
40 implemented, but it requires CAS (code access security) feature. We need to
41 fixup this class after ongoing CAS effort works.
43 You might also be interested in an improved <a href="http://codeblogs.ximian.com/blogs/benm/archives/000039.html">XmlCachingResolver</a> by Ben Maurer.
44 If even one time download is not acceptable, you can use <a href="http://primates.ximian.com/~atsushi/XmlStoredResolver.cs">this one</a>.
46 [2.0] XmlDataSourceResolver is not implemented as yet.
50 NameTable is implemented, but also needs performance improvement.
51 It affects on the whole XML processing performance so much.
52 Optimization hackings are welcome. There is also a <a
53 href="http://bugzilla.ximian.com/show_bug.cgi?id=59537">bugzilla entry</a>
58 XmlTextReader, XmlNodeReader and XmlValidatingReader are almost finished.
61 * All OASIS conformance test passes as Microsoft does. Some
62 W3C tests fail, but it looks better.
63 * Entity expansion and its well-formedness check is incomplete.
64 It incorrectly allows divided content models. It incorrectly
65 treats its Base URI, so some dtd parse fails.
66 * I won't add any XDR support on XmlValidatingReader. (I haven't
67 ever seen XDR used other than Microsoft's BizTalk Server 2000,
68 and Now they have 2002 with XML Schema support). If anyone
69 contributes an implementation, it would be still nice.
72 XmlTextReader and XmlValidatingReader should be faster than now. Currently
73 XmlTextReader looks nearly twice as slow as MS.NET, and XmlValidatingReader
74 (which uses this slow XmlTextReader) looks nearly three times slower. (Note
75 that XmlValidatingReader wouldn't be so slow as itself. It uses schema
76 validating reader and dtd validating reader.)
81 The design of Mono's XmlValidatingReader is radically different from
82 that of Microsoft's implementation. Under MS.NET, DTD content validation
83 engine is in fact simple replacement of XML Schema validation engine.
84 Mono's DTD validation is designed fully separate and does validation
85 as normal XML parser does. For example, Mono allows non-deterministic DTD.
87 Another advantage of this XmlValidatingReader is support for *any* XmlReader.
88 Microsoft supports only XmlTextReader (this bug is fixed in .NET 2.0 beta,
89 taking shape of XmlReader.Create()).
91 <del>I added extra support interface named "IHasXmlParserContext", which is
92 considered in XmlValidatingReader.ResolveEntity(). </del><ins>This is now
93 made as internal interface.</ins> Microsoft failed to design XmlReader
94 so that XmlReader cannot be subtree-pluggable (i.e. wrapping use of other
95 XmlReader) since XmlParserContext shoud be supplied for DTD information
96 support (e.g. entity references cannot be expanded) and namespace manager.
97 (In .NET 2.0, Microsoft also supported similar to IHasXmlParserContext,
98 named IXmlNamespaceResolver, but it still does not provide DTD information.)
100 We also have RELAX NG validating reader (described later).
103 *** System.Xml.Schema
107 Basically it is completed. You can test how current schema validation engine
108 is complete (incomplete) by using standalone test module (see
109 mcs/class/System.XML/Test/System.Xml.Schema/standalone_tests).
110 At least in my box, msxsdtest fails only 30 cases with bugfixed catalog -
111 this score is better than that of Microsoft implementation. But instead,
112 we need performance boost. There should be many points to improve
113 schema compilation and validation.
115 **** Schema Object Model
117 Completed, except for some things to be fixed:
120 * Complete facet support. Currently some of them is missing.
121 Recently David Sheldon is doing several fixes on them.
122 * ContentTypeParticle for pointless xs:choice is incomplete
123 (fixing this arose another bugs in compilation.
124 Interestingly, MS.NET also fails around here, so it might
125 be nature of ContentTypeParticle design)
126 * Some derivation by restriction (DBR) handling is incorrect.
129 **** Validating Reader
131 Basically this is implemented and actually its feature is complete,
132 but I have only did validation feature testing. So we have to write more
133 tests on properties, methods, and events (validation errors).
136 *** System.Xml.Serialization
140 Well, in fact XmlSerializer is almost finished and is on bugfix phase.
142 However, we appliciate more tests. Please try
145 * System.Web.Services to invoke SOAP services.
146 * xsd.exe and wsdl.exe to create classes.
149 And if any problems were found, please file it to bugzilla.
151 Lluis also built interesting standalone test system placed under
152 mcs/class/System.Web.Services/Test/standalone.
154 You might also interested in "genxs", which enables you to create custom
155 XML serializer. This is not included in Microsoft.NET.
157 href="http://primates.ximian.com/~lluis/blog/archives/000120.html">here</a>
158 and manpages for details. Code files are in mcs/tools/genxs.
160 Lluis also created "sgen", that based on XmlSerializer.GenerateSerializer().
161 Code files are in mcs/tools/sgen.
163 *** System.Xml.XPath and System.Xml.Xsl
165 There are two XSLT implementations. One and historical implementation is
166 based on libxslt (aka Unmanaged XSLT). Now we uses fully implemented and
167 managed XSLT by default. To use Unmanaged XSLT, set MONO_UNMANAGED_XSLT
168 environment value (any value is acceptable).
170 As for Managed XSLT, we support msxsl:script.
172 It would be nice if we can support <a href="http://www.exslt.org/">EXSLT</a>.
173 <a href="http://msdn.microsoft.com/WebServices/default.aspx?pull=/library/en-us/dnexxml/html/xml05192003.asp">Microsoft has tried to do some of them</a>,
174 but it is not successful because of System.Xml.Xsl design problem:
177 * In general, .NET's "extension objects" (including
178 msxsl:script) is not useful to return node-sets (MS XSLT
179 implementation rejects just overriden XPathNodeIterator,
180 but accepts only their hidden classes. And are the same
181 in Mono though classes are different)
183 * In .NET's extension object design, extension function name
184 is a valid method name that cannot contain some characters
185 such as '-'. That is, implementing EXSLT in C# is impossible.
188 So if we support EXSLT, it has to be done inside our System.XML.dll.
189 Microsoft developers are also aware of this problem and some of them wish
190 to have EXSLT support in WinFX (not whidbey). If anyone is interested
191 in it, it would be nice.
193 Our managed XSLT implementation is slower than MS XSLT for some kind of
194 stylesheets, and faster for some.
199 I implemented an experimental RelaxngValidatingReader. It is still not
200 complete, for example some simplification stuff (see RELAX NG spec
201 chapter 4; especially 4.17-19) and some constraints (especially 7.3).
202 See mcs/class/Commons.Xml.Relaxng/README for details.
207 * Custom datatype support. Right now, you can use XML schema
208 datatypes ( http://www.w3.org/2001/XMLSchema-datatypes )
209 as well as RELAX NG default datatypes (as used in relaxng.rng).
211 * RELAX NG Compact Syntax support, though not yet stable.
212 See Commons.Xml.Relaxng.Rnc.RncParser class.
218 Microsoft released the first public beta version of .NET Framework 2.0,
219 available from <a href="http://www.microsoft.com/downloads/details.aspx?familyid=916EC067-8BDC-4737-9430-6CEC9667655C&displaylang=en">MSDN</a>.
220 It contains several new classes.
222 There are two assemblies related to System.Xml v2.0; System.Xml.dll and
223 System.Data.SqlXml.dll. Now that System.Data.SqlXml.dll is little important.
224 It just contains only XQueryCommand class inside System.Xml.* namespace.
225 Most of the important part are in System.Xml.dll.
227 Note that .NET Framework is pre-release version, so they are subject
228 to change. Actually many of the pre-released classes vanished.
230 System.Xml 2.0 contains several features such as:
233 * new XPathNavigator <del>and XPathDocument</del><ins>XPathDocument is <a href="http://blogs.msdn.com/dareobasanjo/archive/2004/08/25/220251.aspx">being reverted</a></ins>
234 * XmlReaderSettings, XmlWriterSettings and factory methods
235 * Strongly typed XmlReader and XmlWriter.
236 * XML Schema design changes
238 * Well-documented and improved XmlSerializer.
239 * XQuery execution engine
240 * XQuery and XSLT per-stylesheet assembly generator
245 **** XmlReader/XmlWrier Factory methods
247 In .NET 2.0, XmlTextReader, XmlNodeReader, XmlValidatingReader are
248 obsolete and XmlReader.Create() is recommended (there is however no
249 alternative way to create XmlNodeReader). Similarly, there are
250 XmlWriter.Create() overloads.
252 Currently, Microsoft's XmlWriter.Create() is unreliable and maybe there
253 will be changes. So basically XmlWriter.Create() is supposed to be done
254 after the next beta version of .NET 2.0.
256 Some of XmlReader.Create() overloads are implemented, with limited
257 XmlReaderSettings support.
260 **** Typed XmlReader/XmlWriter
262 In .NET 2.0, XmlReader is supposed to support strongly-typed data reading.
263 They are based on W3C "XML Schema Datatypes" Recommendation and "XQuery 1.0
264 and XPath 2.0 Data Model" Working Draft.
266 Some of XmlReader.ReadValueAsXxx() and XmlWriter.WriteValue() overloads are
267 implemented, though incompletely. They are based on internal XQueryConvert.
270 **** Sub-tree handling in XmlReader/XmlWriter/XPathNavigator
272 Currently XmlReader.ReadSubtree(), XmlWriter.WriteSubtree() and
273 XPathNavigator.ReadSubtree() are implemented, though not well-tested.
274 They are based on Mono.Xml.SubtreeXmlReader and
275 Mono.Xml.XPath.XPathNavigatorReader classes.
278 *** System.Xml.Schema 2.0
280 Since .NET 1.x is not so compliant with W3C XML Schema specification,
281 Microsoft had to redesign System.Xml.Schema classes. We also have to
284 1) It does not expose XmlSchemaDatatype anymore (except for obsolete
285 members). Primitive types are represented as XmlSchemaSimpleType
286 instances (thus there are ElementSchemaType, AttributeSchemaType,
287 BaseXmlSchemaType that replace some existing properties).
289 2) "XQuery 1.0 and XPath 2.0 Data Model" datatypes (such as
290 xdt:dayTimeDuration) are newly supported. They are partially implemented
291 yet. This task is partly done.
293 3) schema structures are now bound in parent-child relationship. It is
294 not yet implemented. With related to it, there seems bunch of schema
295 compilation bugfixes.
297 4) XmlSchemaCollection is not used anymore to represent effective set of
298 schemas. Instead, new XmlSchemaSet class is used. It should affect on
299 schema compilation design. In fact, I've implemented XmlSchemaCollection
300 as more conformant to W3C specification, but there are still many changes
301 required. This task is partly done.
306 In .NET 2.0, there is an XML Schema inference implementation. Now that
307 XmlSchemaSet is basically implemented, it can be separately done by anyone.
308 Volunteer efforts are welcome here.
311 *** System.Xml.XPath 2.0
313 **** Editable XPathDocument
316 in .NET 2.0 XPathDocument is supposed to be editable. Currently we provide
317 fast document table model based implementation (DTMXPathNavigator), but
318 by that design change, we (and they) cannot provide fast read only
319 XPathNavigator from XPathDocument anymore.
321 It is being reverted to the original (.NET 1.x) XPathDocument. We still have
322 them, but we'll revert them too in the future. So our XPathDocument will be still faster one.
325 Currently, new XPathDocument implementation is provided. The actual
326 implementation is Mono.Xml.XPath.XPathDocument2, that is simple dom-like
327 tree model. XPathDocument2 implements the same interfaces as XPathDocument
328 does. And XPathDocument delegates most of the methods to that class (for
329 example, XPathDocument.CreateEditor() calls XPathDocument2.CreateEditor()).
331 Currently Mono.Xml.XPath.XPathDocument2 is unstable (it does not pass
332 the standalone XSLT tests unlike existing DTMXPathDocument does). So
333 it did not replace existing XPathDocument implementation, but you can use
334 new implementation by explicitly setting environment value
335 USE_XPATH_DOCUMENT_2 = yes. Currently it supports (well, is supposed
336 to support) basic editor feature such as AppendChild(). Other members
337 are untested (such as RejectChanges()).
339 **** extra stuff - XPathEditableDocument
341 Currently we provide another IXPathEditable; XPathEditableDocument. That is
342 based on the idea that handles XmlDocument as editor target. It is
343 implemented as Mono.Xml.XPath.XPathEditableDocument. We might provide this
344 class as extra set (might be different mono-specific XML assembly).
347 **** System.Xml.XQuery
349 In this namespace, there are two significant classes. XsltCommand and
352 XsltCommand implements XSLT transformation. It is almost the same as
353 System.Xml.Xsl.XslTransform, but this class transforms documents twice
354 to four times as fast as XslTransform. Instead, stylesheet compilation
355 is much slower, because it generates compiled stylesheet assembly.
357 XQueryCommand implements XQuery. XQuery is a new face XML document
358 manipulation language (at least new face in .NET world). It is similar
359 to XSLT, but extended to support XML Schema based datatypes (and it is
360 not XML based langauge). It is similar to XPath, but it can construct
361 XML nodes. It has no complicated template resolution, but works like
362 functional languages.
364 Under MS.NET, XQuery implementation is mainly in System.Xml.Query and
365 MS.Internal.Xml.* namespaces. The implementation is mostly
366 in System.Xml.dll. It is also true to our System.Xml.dll. Our XQueryCommand
367 in System.Data.SqlXml.dll just invokes the actual XQuery processor
368 (Mono.Xml.XPath2.XQueryCommandImpl) which resides in System.Xml.dll via
371 Currently we are not implementing MS.Internal.Xml.* classes. MS
372 implementation is based on an old version of the W3C specification, and
373 our implementation is currently based on
374 <a href="http://www.w3.org/TR/2004/WD-xquery-20040723/">23 July 2004
375 version</a> (latest as of now) of the working draft.
377 XQuery implementation tasks are:
380 * XQuery syntax parser that parses xquery string to AST
381 (abstract syntax tree). -> partly not done.
383 * XQuery AST compiler into static context -> partly not done.
385 * XQuery (dynamic context) runtime = XQuery expression evaluator
386 + sequence iterator. -> partly not done.
388 * XPathItem data model and (mainly) conversion support.
391 * Applied expression classes for XQuery/XPath 2.0 functions and
392 operators. -> partly done.
394 * Optimization, and design per-query assembly code generator (later)
397 It already handles some queries, while major part implementation is missing
398 or buggy (like FLWOR, expressions for sequence type handling, built-in
399 function support etc.).
402 *** Relax NG and DSDL in Mono 1.2
404 Currently we support only RELAX NG as one part of ISO DSDL effort. There
405 is existing Schematron implementation (NMatrix Project: <a
406 href="http://sourceforge.net/projects/dotnetopensrc/">
407 http://sourceforge.net/projects/dotnetopensrc/</a>). With a few changes,
408 it can be used with mono.
410 We also don't have multi-language based validation support, namely
411 Namespace-based Validation Dispatch Language (NVDL). To support unwrapping,
412 one special XmlReader implementation is required (other schema validation
413 support can be done by ReadSubtree()). Note that we had seen RELAX
414 Namespace, Modular Namespace (MNS) and Namespace Routing Language (NRL)
415 - that is, standardization effort is still ongoing (though NVDL looks
416 mostly the same as NRL).
418 In Mono 1.2, there might be improvements on Commons.Xml.Relaxng.
421 * Currently RelaxngPattern.Compile() provides cheap compilation
422 error information. At least it can provide error location.
423 Also, the type of error should be kind of
424 RelaxngGrammarException.
426 * Right now there is no ambiguity detection implementation that
427 would be useful for RelaxngPattern based xml serialization (if
430 * Because of lack of ambiguity detection, there is no way to
431 provide XmlMapping (XmlTypeMapping/XmlMemberMapping). But
432 If anyone is interested in such effort, integration with
433 XmlSerializer would be interesting task.
441 See <a href="ado-net.html">ADO.NET page</a>.
446 *** Mutual assembly dependency
448 Sometimes I hear complain about System.dll and System.Xml.dll mutual
449 dependency: System.dll references to System.Xml.dll (e.g.
450 System.Configuration.ConfigXmlDocument extended from XmlDocument), while
451 System.Xml.dll vice versa (e.g. XmlUrlResolver.ResolveUri takes System.Uri).
452 Since they are in public method signatures, so at least we cannot get rid
453 of these mutual references.
455 Nowadays System.Xml.dll is built using incomplete System.dll (lacking
456 System.Xml dependent classes such as ConfigXmlDocument). Full System.dll
457 is built after System.Xml.dll is done.
459 Note that you still need System.dll to run mcs.
462 Atsushi Eno <asushi@ximian.com>
463 last updated 09/02/2004