mirror of
https://github.com/BepInEx/mono.git
synced 2026-11-16 17:10:20 -10:00
Originally this implementation checked recursion based on RELAX NG-inspired algorithm explained at http://www.thaiopensource.com/relaxng/simplify.html. This works only for "simple" syntax, which premises that every <element> is defined under top-level <define> element. XML Schema doesn't follow this and could result in infinite loop like reported. The latest REC-xmlschema-1 denotes the explicit rule on illegal recursion on 3.8.6 (Model Group Correct) as: 2 Circular groups are disallowed. That is, within the {particles} of a group there must not be at any depth a particle whose {term} is the group itself. ... so, I simply replaced the checker with simple Stack. (oh, and for old Stack I haven't bothered to replace non-generic collection everywhere. XSD is as dead as System.Collection.)
* System.Xml.Schema standalone test collections ** Small XML Schema validation test system:: This is a small standalone test system using W3C XML Schema test collection. http://www.w3.org/2001/05/xmlschema-test-collection.html "make" will prepare whatever you need to run test (downloading test archive from W3C, expanding, compiling test runner). "make run-test" will run the tests and reports as TestResult.xml. ** Post Schema Compilation Information comparison test:: This is a test for Post Schema Compilation Information properties. xs-pcsi-compare.exe emits properties of schema object model. pcsi-ms-all.zip contains msxsdtest result I got from MS.NET 1.1. (Currently no other way to compare the results by hands.) As for the result differences, some notes would be required: - MS.NET holds incorrect base schema type for simple list and union types. Their base type is simple ur-type (3.14.7) but MS.NET holds item type(s) there. - For unspecified types for element and attribute, they incorrectly specifies xs:string type. However in those cases, attribute types should be mapped to simple ur-type and element types should be (complex) ur-type (3.4.7) This will be added in the future (currently Mono has the same problem). - MS.NET does not remove "pointless" choice particle (see 3.9.6). However it results in incorrect validation. (e.g. msxsdtest/Particles/particlesHb010.xsd) Atsushi Eno <[email protected]>