-Bugs in Implementation:\r
-\r
-1. XmlSchema.Read Fails if there is a wrong value for a enum type. Eg form="none" will throw an exception.\r
- The only conformant solution to this is using our own Deserializer or parser.\r
-2. Schema Compile ignores errors where a name or id (or any other attribute) is empty. Eg if a schama has name="",\r
- it will not throw an exception but will be silently ignored. Again the solution is to write own parser.\r
-3. Not all Compile methods check for the Id attribute. Trivial but needs to be done.\r
-4. Need to check if errorCount is calculted correctly, and all of the error() methods increase it.\r
+Bugs in Mono Implementation of XmlSchema:\r
+\r
+1.\r
+None of the XmlSchemaObjects except the XmlSchema populate XmlSerializerNamespaces.\r
+\r
+2.\r
+Non schema attributes are not being handled. Need an internal XmlAttribute constructor.\r
+3.\r
+Documentation and appInfo's Markup is not being read in the Read() Method.\r
+\r
+4.\r
+Handling of unique particle attribution of sequences is almost complete, \r
+but not perfect. Note that it is W3C specification that is so buggy. \r
+Basically there is neither formalization nor even normative definition \r
+about unique particle attribution. Appendix H tried to describe about \r
+that, but it is too incomplete.\r
+\r
+5.\r
+Currently ContentTypeParticle does not eliminate "pointless" (see 3.9.6 of\r
+xml schema structures spec) XmlSchemaChoice, like MS.NET. It causes\r
+incorrect DBR validation.\r
+\r
+6.\r
+See MS-BUGS.txt 008 and 014. There are some correction and we should fix our\r
+implementation based on those wrong premises.\r
+\r
projects / mono.git / blobdiff