On Jan 9, 2005, at 7:24 AM, Ceki G�lc� wrote:
Hello Curt,
I see that you have started work on XML Schema for http://logging.apache.org/.
Here are some initial comments.
- latest CVS no longer includes support for error handlers, so including them in the Schema does not make sense.
I'll let you remove the references to errorHandler in the Schema.
This pass at the schema was just driven by the content of input/src/xml with no reference to the code. tests/input/xml/fallback1.xml contains an errorHandler element and it appears to be used by o.a.l.varia.ErrorHandlerTestCase. Is that test dead or just that DOMConfigurator still supports it but Joran won't?
I toyed with the idea of have both a strict and lax schema. The current schema is stricter than the code base. For example, it specifies a specific order to the elements, but that is necessary to enforce multiplicity (such as an appender only having 0 or 1 layouts). A lax schema would allow you to not follow a fixed order, but would not catch missing or extra elements. Having a strict schema is more helpful for schema-driven editors since it gives you less choices on the next step. If we decide that we want to have a lax schema, it could be generated from the strict schema.
- If you can, please make an effort to conform to the log4j indentation style. You can find log4j's coding conventions expressed in Jalopy under the file ./src/log4j-coding-convention.xml.
Okay, I'll review it. I used Checkstyle in the past but not Jalopy. I'm guessing from some artifacts in some of your recent files that you use Eclipse (which I what I used for that code). Any hints on configuring Eclipse for a compatible style?
- The org.apache.log4j.xml.SchemaTestCase requires Xerces. For a mysterious reason, the latest gump run complains about the unavailability of Xerces even of it is part of the project description for log4j-tests. Go figure.
Okay, unless you beat me to it. I'll remove Schema from the regression target.
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
