Saxonica Developer Community

Today

• 
  Michael Kay Saxon

  13:01 · Commented on issue

  well, you can always supply data directly from the calling application to the class implementing the extension function before the Saxon transformation starts... Bug #4434: Using SaxonTransformerFactory sometimes leaves the Configuration without an APIProvider

  Hello,
  For me it all starts with SaxonTransformerFactory:
  1. The default constructor creates an Processor which in turn calls net.sf.saxon.Configuration.setProcessor(ApiProvider). This means that an Initializer instance will have acc...
• 
  Michael Kay It's legitimate for the ApiProvider to be null. Saxon itself appears to no longer use this property. I think it was provided primarily for extension functions, for example if we supply a node to an extension function then the extension f...

  11:15 · Commented on issue
• 
  Michael Kay Saxon

  12:31 · Commented on issue

  I think it would be useful to establish whether the source tree is the same in the failing environment from the non-failing environment. It's not particularly easy to do this with XQJ, but it can be done like this. I'd be grateful if you... Bug #4433: Saxon 9.9.1-5 outputs malformed XML due to duplicate default namespace declaration

  This bug occurs with Saxon 9.9.1-5 and -6 using the XQJ API. (I have previously also mention it on [saxon-help](https://sourceforge.net/p/saxon/mailman/message/36881831/), but unfortunately got no response there.)
  I have attached a si...
• 
  Michael Kay These notes capture my investigation and are largely for my own benefit.
  Under XQJ we start evaluation in pull mode using `XQueryExpression.iterator()` which calls `expression.iterate(`). The function call to `local:atom-entry-content...

  12:12 · Commented on issue
• 
  Michael Kay I have run using the JDK parser and I'm still not seeing the problem. It's a slightly different version of the JDK - jdk1.8.0_121.jdk on Mac OS, but that doesn't seem likely to be a signficant difference.
  I'm now studying exactly how ...

  11:06 · Commented on issue
• 
  Michael Kay OK, I'll try running with the built-in JDK parser and see if that's the significant difference.
  In the past I always recommended running with the Apache parser because the JDK parser had some serious bugs and Oracle seemed to be doing...

  09:48 · Commented on issue
• 
  Michael Kay As to the question, why is XQJ different from the command line, there can be a number of possibilities. Both environments will be setting up a pipeline containing parser, query engine, and serializer, but there can be many differences in...

  01:43 · Commented on issue
• 
  Michael Kay In my first attempt to run this, I can't reproduce the bug. I'm getting the same output with XQJ as with the command line (that is, no redundant namespace declarations).
  The fact that you're seeing incorrect output in 9.9.1.4 as well ...

  01:22 · Commented on issue

2020-01-19

• 
  Michael Kay Saxon

  00:15 · Edited issue

  Bug #4433 (In Progress): Saxon 9.9.1-5 outputs malformed XML due to duplicate default namespace d...

  This bug occurs with Saxon 9.9.1-5 and -6 using the XQJ API. (I have previously also mention it on [saxon-help](https://sourceforge.net/p/saxon/mailman/message/36881831/), but unfortunately got no response there.)
  I have attached a si...

2020-01-17

• 
  Michael Kay Saxon

  23:26 · Closed issue

  Support #4392 (Closed): S9API and JSON input

  I am using the S9API to evaluate XPath expressions, like this:
  Processor processor = new Processor();
  XPathCompiler compiler = processor.newXPathCompiler();
  XPathExecutable executable = compiler.compile(XPATH_EXPRESSION);
  XPath...