Saxonica Developer Community

2020-07-11

• 
  Michael Kay Saxon

  23:08 · Commented on issue

  I guess it's probably a loop-lifting problem - the expression on the RHS being moved out of the implicit loop defined by the "/" operator because it has no functional dependency on the value of the expression on the left. Bug #4640: Reference to a variable as an XPath step

  In Saxon9HE 9.8.0.15 this expression works fine using the first XPath step as an "if" statement:
  ~~~ xml
  <xsl:copy-of
  select="self::table-wrap[ancestor::fig]/$c:f/c:figTableCaptionFonts/@*"/>
  ~~~
  In Saxon9HE 9.9.1.7 the ...
• 
  Michael Kay I'm wondering a little bit what this code is trying to do. Using a variable reference on the RHS of "/" is legitimate, but it's not very usual.
  ~~~
  self::table-wrap[ancestor::fig]/$c:f/c:figTableCaptionFonts/@*
  ~~~
  The semantics ...

  20:25 · Commented on issue
• 
  Michael Kay You could try and see if it's an optimization problem by running with optimizations disabled - use the -opt flag on the command line. To drill down, you can suppress optimizations selectively (function inlining and loop lifting are the m...

  20:13 · Commented on issue
• 
  Michael Kay I'm afraid it's impossible to diagnose this kind of issue unless there's something I can run. A small repro is ideal, but if you can't produce that then a big complicated repro will do. Without a repro (especially for cases of "wrong res...

  19:33 · Commented on issue

2020-07-10

• 
  Michael Kay Saxon-JS

  09:24 · Commented on issue

  Could you expand on this please? Bug #4639: SaxonJS in Electron appears non-standard.

2020-07-09

• 
  Michael Kay Saxon

  13:35 · Commented on issue

  Fixing the code in optimizeValueComparison has exposed another bug in code which previously wasn't being executed. The symptom is
  ~~~
  java.lang.NullPointerException
  at net.sf.saxon.expr.CompareToIntegerConstant.effectiveBooleanValu... Bug #4610: Bug in XPath optimizer with positional filter

  Hello,
  I'm using the XPath engine of Saxon HE 9.8 (JRE). The optimizer fails to optimize the following form of expression
  ~~~ xml
  //A[true() = false()]/*[2]
  ~~~
  with this exception
  ~~~
  net.sf.saxon.trans.XPathException: Axis...
• 
  Michael Kay Saxon/C

  12:34 · Commented on issue

  First thing to do is to check that it's actually Saxon doing both (a) the serialization, and (b) the encoding. That's going to depend on which APIs you are using to invoke the transformation. If the API returns the result as a string rat... Support #4638: Output xsl3 transform to UTF-16LE with BOM

  Hi all,
  I'm testing Saxon/C with C++ (version PE). While transform without <xsl:output> create file encoded with UTF-8, trying to output to UTF-16LE (windows) with <xsl:output encoding="utf-16le" byte-order-mark="yes"/> create output en...
• 
  Michael Kay Saxon

  10:41 · Commented on issue

  I think this is what Logger.info() was intended for; but we aren't really using the mechanism properly: we have almost no code that changes the logging threshold, and we have other mechanisms in parallel, e.g. the -t flag doesn't change ... Feature #4637: Support a .debug() or similar method on Logger

  It would be nice to use the Saxon Logger interface, rather than dragging in a(nother) logging framework to print diagnostic messages. But there's no method on Logger that's usually below the threshold at which messages are printed. Consi...

2020-07-08

• 
  Michael Kay Saxon

  15:55 · Closed issue

  I have replaced the CONSTRAINED_CLASS property of `OperandRole` with a property that describes an explicit constraint on the expression participating in that role; for patterns the constraint is that the expression must be a pattern, it ... Bug #4636 (Resolved): AssertionError compiling xsl:for-each-group with a higher-order-function

  A compile-time assertion error occurs when compiling an `xsl:for-each-group` instruction with `group-starting-with="*[$starts-function(.)]`, where $starts-function is an anonymous inline function.
• 
  Michael Kay What's happing is that optimization of the function els:wrap-elements-starting-with-names attempts to inline the function els:wrap-elements-starting-with, and after inlining, the reoptimization of the xsl:for-each-group function now know...

  12:33 · Commented on issue