New RNC test suite
"Jeff Rafter" <[email protected]> Fri, 14 May 2004 01:20:06 -0700
| Newsgroups | gmane.text.xml.relaxng.general |
|---|---|
| Message-ID | <011101c4398c$44149220$6403a8c0@ARIMATHEA> |
Well, my question about an RNC test suite went unanswered last week, so in my need, I began one. I am trying to be very thorough. I broke down the tests into various categories: encoding, tokenization, grammar, constraints. I completed a set of tests for tokenization. There are 120 tests which include illegal constructions and legal constructions. Some of the tests are redundant checking the same type of thing in various productions. The tests can be found (currently) at http://www.jeffrafter.com/anglia/anglia-test-suite.zip This will almost definitely change (probably to a CVS repository somewhere). "Anglia" is the working name for the RNC toolset I am creating*. Included in the zip file is anglia.exe which can run the tests (and check other RNC files). It passes this group of tests though there are quite a few other sections of code that are commented out right now as I work on things. It should work in Mono and Microsoft.NET anglia -t test/RNCTest.xml Please send any comments, though I am really looking for comments on the tests more than the parser at this point. I have spot checked Jing and XMLDistilled and found a couple of errors in the latter-- these may be problems with the tests themselves. * Because the parser is based on the methodology of the AElfred parser (which is also used in "Saxon"), I wanted to choose a name in this set-- also I was looking for a name with "ng" in it. Apart from these two factors "Anglia" sounds a lot like "angle" which may hint at its use with XML files and their <angle> brackets. Or it may not. All the best, Jeff Rafter