Saxonica Developer Community

2017-09-22

• 
  Michael Kay Saxon

  23:51 · Closed issue

  Rewrote the EquivalenceComparer used for the XQuery switch expression. It now inherits from GenericAtomicComparer rather than AtomicSortComparer. It includes the logic to accept a run-time dynamic context for use when the comparison depe... Bug #3457 (Resolved): switch, xs:QName

  With Saxon-EE 9.7.0.8J, the following query returns 2 instead of 1:...
• 
  Michael Kay Thanks for reporting it. Looking first at the QName case:
  I have made this into QT3 test case switch-014 (yes, the test number gives a clue that the test coverage for this new feature is somewhat minimal...)
  The switch expression c...

  20:55 · Commented on issue
• 
  Michael Kay Saxon

  23:36 · Replied in forum

  Yes, this shouldn't normally happen during XSLT processing (duplicate attributes cause an error in XQuery, but on XSLT the last one wins).
  I'm not entirely surprised that it should happen, however, because the mechanism by which the S... Help: RE: Saxon-EE 9.7.0.20: seeing error message XQDY0025

  We have upgraded Saxon-EE from 9.6 to 9.7.0.20. Then while performing XSLT transform, we started bumping into this fatal error: ...
• 
  Michael Kay Saxon

  23:21 · Replied in forum

  Logged as a bug here:
  https://saxonica.plan.io/issues/3458?issue_count=31&issue_position=1&next_issue_id=3457 Help: RE: Memory problems after migration from 9.1.0.8 to 9.8.0.4

  Hello,
  I'm currently trying to migrate from Saxon 9.1.0.8 to 9.8.0.4 PE (evaluation licence for development and tryout) in my copmany java application and I have encountered memory problems.
  First thing is that we are caching XQuer...
• 
  Michael Kay Thanks for reporting it and for your useful analysis of the cause of the problem.
  I may have to ask for a repro (code I can run to reproduce the problem) but I'll try and see if I can trigger it from the information provided. I can se...

  20:58 · Replied in forum
• 
  Michael Kay Saxon

  23:21 · Closed issue

  Although I don't have a repro that actually causes a SameNameTest to be placed in the TypeHierarchy cache, it's clearly something that shouldn't be allowed to happen so I have put in a guard against it. Bug #3458 (Resolved): Putting SameNameTest in TypeHierarchy cache locks document into memory

  A SameNameTest is a pattern than matches nodes having the same node kind and name as a specified node. Because this object contains a reference to the relevant NodeInfo, storing it in the TypeHierarchy cache will result in the document t...
• 
  Michael Kay A SameNameTest is a pattern than matches nodes having the same node kind and name as a specified node. Because this object contains a reference to the relevant NodeInfo, storing it in the TypeHierarchy cache will result in the document t...

  23:19 · Created issue

2017-09-21

• 
  Michael Kay Saxon-JS

  23:30 · Replied in forum

  I'm very keen to deliver this, but I'm afraid it's too early to give any idea of dates. It's not imminent, I'm afraid. General: RE: Timeline for Node.js support?

  In the announcement blog, there was a mention a while back that support for Node.js would be on the roadmap -> https://saxonica.plan.io/projects/saxon-js/news. Any updates on this?
  Thanks
• 
  Michael Kay Saxon

  12:26 · Commented on issue

  We're going to be putting together a maintenance release next week. It usually takes a couple of days, but it depends on what problems we find in testing. Yes, I'll consider putting out the bulk copy under a configuration flag. Bug #3443: Performance degradation between 9.2 and 9.8

  A user has submitted an XSLT transform that takes 102s under 9.2, increasing to 241s under 9.8.
  *Summary: the main problem was that with hotspot bytecode generation enabled (the default setting for Saxon-EE), a recursive function that...

2017-09-20

• 
  Michael Kay Saxon

  13:39 · Commented on issue

  I have made the change so the body of a function is always compiled with iterate() and process() methods. (An evaluateItem() or effectiveBooleanValue() method will be generated if the function is called in that mode).
  The number of tr... Bug #3443: Performance degradation between 9.2 and 9.8

  A user has submitted an XSLT transform that takes 102s under 9.2, increasing to 241s under 9.8.
  *Summary: the main problem was that with hotspot bytecode generation enabled (the default setting for Saxon-EE), a recursive function that...