From: Kenneth Reid Beesley <krbeesley(_at_)gmail(_dot_)com>
Subject: RE: [xsl] How to copy attribute value to text? (Suspected bug
involving supplementary characters)
Date: 7 July 2016 at 12:23:29 MDT
To: xslt <xsl-list(_at_)lists(_dot_)mulberrytech(_dot_)com>
***** Suspected bug involving supplementary characters *****
But my real task involves an input XML document, in UTF-8 encoding, that
consists of Deseret Alphabet characters, which are encoded in the
supplementary area. In such a case, the resulting text content in the <word>
element, copied from an original attribute value, is corrupted. I saw such
corruption in my own attempts, and couldn’t understand what was happening.
Using the following input document (the Deseret Alphabet characters may not
display correctly for you)
<?xml version="1.0" encoding="UTF-8"?>
<foo>
<bar>𐑄𐐮𐑅 𐐮𐑆 𐐾𐐲𐑅𐐻 <word correction="𐐻𐐭">𐑂𐐯𐑉𐐮</word> 𐑁𐐲𐑌𐐮</bar>
</foo>
the output, using your script, is corrupted. The text() value in the output
is not the same as the original @correction value. Extra characters (just
one in this case) are inserted. The longer the original attribute value, the
more extra characters are inserted.
<?xml version="1.0" encoding="UTF-8"?>
<foo>
<bar>𐑄𐐮𐑅 𐐮𐑆 𐐾𐐲𐑅𐐻 <word origerror="𐑂𐐯𐑉𐐮">𐐻𐐻𐐭</word> 𐑁𐐲𐑌𐐮</bar>
</foo>
This kind of corruption is exactly what I was seeing using my own scripts,
leading me to bother the group.
I suspect a bug in the XSLT engine involving supplementary characters.
Again, I’m using SaxonHE9-7-0-6J.
What’s my next step?
Thanks,
Ken
From: Michael Müller-Hillebrand <mmh(_at_)docufy(_dot_)de>
Subject: Re: [xsl] How to copy attribute value to text? (Suspected bug
involving supplementary characters)
Date: 7 July 2016 at 14:20:30 MDT
To: xsl-list(_at_)lists(_dot_)mulberrytech(_dot_)com
When copying the data and stylesheet into OxygenXML and also enabling bidi
support, the XSLT processing works fine.
<?xml version="1.0" encoding="UTF-8"?>
<foo>
<bar>𐑄𐐮𐑅 𐐮𐑆 𐐾𐐲𐑅𐐻 <word origerror="𐑂𐐯𐑉𐐮">𐐻𐐭</word> 𐑁𐐲𐑌𐐮</bar>
</foo>
So your problems may come form some details in your setup? How are you
running the transform?
BTW, interesting letters!
- Michael
I _was_ running the transform with the default JDK XML parser (Java 1.8). I’m
using SaxonHE9-7-0-6J.
This default JDK parser is reputed to be buggy.
From: Michael Kay <mike(_at_)saxonica(_dot_)com>
Subject: Re: [xsl] How to copy attribute value to text? (Suspected bug
involving supplementary characters)
More likely to be a bug in the JDK parser. Try it using Apache Xerces, which
is much more reliable than the JDK parser. I think some of the long-standing
bugs in the JDK parser have finally been fixed in Java 8, so you could also
try it with a different JDK.
Michael Kay
Saxonica
Michael Kay is right. I changed to using the Xerces-J parser and now
everything works as expected.
By the way, I found it a little difficult to figure out how to use Saxon and
specify the xerces parser.
I had to hunt around a bit. I finally found the following incantation (as
coded in my Makefile).
# using Saxon XSLT with the Xerces-J parser
BoMDA1869c.xml: BoMDA1869.xml BoMDA1869c.xsl
java
-Djavax.xml.parsers.DocumentBuilderFactory=org.apache.xerces.jaxp.DocumentBuilderFactoryImpl
\
-Djavax.xml.parsers.SAXParserFactory=org.apache.xerces.jaxp.SAXParserFactoryImpl
\
net.sf.saxon.Transform -o:$@ $< BoMDA1869c.xsl
I have saxon9he.jar and xercesImpl.jar on my CLASSPATH. It all seems to work.
Am I missing anything?
Many thanks to all who responded to my question.
Ken
********************************
Kenneth R. Beesley, D.Phil.
PO Box 540475
North Salt Lake UT 84054
USA
--~----------------------------------------------------------------
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
EasyUnsubscribe: http://lists.mulberrytech.com/unsub/xsl-list/1167547
or by email: xsl-list-unsub(_at_)lists(_dot_)mulberrytech(_dot_)com
--~--