Re: Jing api
Bob Foster <[email protected]> Thu, 13 Jan 2005 11:02:56 -0800
| Newsgroups | gmane.text.xml.relaxng.general |
|---|---|
| Message-ID | <[email protected]> |
Andrew Welch wrote: >>If you must, you can invoke Jing by reading how JC does it and doing >>likewise. The following method is an example. It will either check a >>schema and/or use it to validate an instance. Please excuse >>the mystery >>classes; if their function isn't apparent, ask. [example snipped] > > Thanks for the example - that's exactly what I needed, but I'm puzzled > by 'if you must' and the 'awful kludge' comment: aren't we meant to use > the api? How else does everyone validate against RNG in their apps? As you know, Kohsuke and I and no doubt others have found ways of invoking RNG at the object level, but I'm not sure the ways amount to an API. An API is really more a matter of intent, so you'd have to ask James Clark if any of the ways of calling jing internally are intended to be API. One of the issues, IIRC, is that setting an entityresolver doesn't work 100% in jing because of some classloading tricks it plays. I had to hack the code. If what I was calling was API, I'd have a reasonable expectation such problems would be fixed. Bob > This is a strange stumbling block on the road of RNG elightenment, up > until now everything has been fantastic...