JDK packaged JAXP transformer indent no longer working in OpenJDK 11

[Niels Bertram]

Hi Frank,

the validity of the produced XML is not impaired. But given that we process
and package the schema docs for humans, the readability is as important as
validity. As a workaround, we have been able to rectify this issue by
overriding the internal transformers with the official Xalan 2.7.2 ones.
But given I am not the only one facing this issue, looks like a good case
to evaluate if the changes applied in JDK-8087303 should be considered a
regression. Certainly broke the the functionality we are used to when
specifying indent=yes on the transformer factory.

Cheers,
Niels


On Wed, May 8, 2019 at 4:23 PM Frank Yuan <frank.yuan at oracle.com> wrote:

> Hi Niels
>
> Yes, it's caused by JDK-8087303. But it's not a regression.
>
> This change has no impact to any xml parser at all, although it is not
> backward-compatible for the human readability,
>
> Actually JDK-8087303 tried to make the form better, but the serializer
> can't distinguish whether the space is a meaningful content or just an
> indentation made by the author, so we had to regard the spaces as xml data
> content.
>
> I think your program should not depend on any specified data format, so
> you have not any impact indeed, don't you?
>
> Thanks
> Frank
>
> > Subject: JDK packaged JAXP transformer indent no longer working in
> OpenJDK 11
> >
> > Hi JDK devs,
> >
> > we recently started to upgrade from OpenJDK 8 to 11 and noticed that the
> > Xalan transformer shipped with the JDK is not handling whitespace
> > indentation properly anymore. Whilst it is not impacting the validity of
> > the XML produced, it is rather annoying that indented results now have
> > extra line breaks. E.g. we post-process some generated XSD files for
> > publication but when done in OpenJDK 11 they look horrible.
> >
> > There are a few people seeing this issue:
> >
> >
> https://stackoverflow.com/questions/53596202/javax-xml-transform-transformer-line-endings-no-longer-respect-system-property
> >
> This is not same as your issue.
>
> > https://github.com/CodeFX-org/java-9-wtf/tree/master/xml-transformer
> This talks about JDK-8087303. These changes were described in JDK 9
> release notes.
>
> >
> > It may be related to the changes made in this ticket:
> >
> > https://bugs.openjdk.java.net/browse/JDK-8087303
> >
> > Would it be possible for someone with permissions to raise a defect for
> > this?
> >
> > $ java -version
> > openjdk version "1.8.0_212"
> > OpenJDK Runtime Environment (build 1.8.0_212-8u212-b01-1~deb9u1-b01)
> > OpenJDK 64-Bit Server VM (build 25.212-b01, mixed mode)
> >
> > (sources at the end of mail)
> >
> > $ javac Transform.java
> >
> > $ java Transform
> > <?xml version="1.0" encoding="UTF-8"?><books>
> >   <book>
> >     <title>test book</title>
> >   </book>
> > </books>
> >
> > $ java Transform --indent
> > <?xml version="1.0" encoding="UTF-8"?><books>
> >   <book>
> >     <title>test book</title>
> >   </book>
> > </books>
> >
> >
> > $ java -version
> > openjdk version "11.0.3" 2019-04-16
> > OpenJDK Runtime Environment (build 11.0.3+1-Debian-1bpo91)
> > OpenJDK 64-Bit Server VM (build 11.0.3+1-Debian-1bpo91, mixed mode,
> sharing)
> >
> > $ java Transform
> > <?xml version="1.0" encoding="UTF-8"?><books>
> >   <book>
> >     <title>test book</title>
> >   </book>
> > </books>
> >
> > $ $ java -Djaxp.debug=1 Transform --indent
> > JAXP: find factoryId =javax.xml.transform.TransformerFactory
> > JAXP: loaded from fallback value:
> > com.sun.org.apache.xalan.internal.xsltc.trax.TransformerFactoryImpl
> > JAXP: created new instance of class
> > com.sun.org.apache.xalan.internal.xsltc.trax.TransformerFactoryImpl using
> > ClassLoader: null
> > <?xml version="1.0" encoding="UTF-8"?><books>
> >
> >     <book>
> >
> >         <title>test book</title>
> >
> >     </book>
> >
> > </books>
> >
> >
> > Source for Transform.java
> >
> > -------------------------------
> >
> > import javax.xml.transform.OutputKeys;
> > import javax.xml.transform.Templates;
> > import javax.xml.transform.Transformer;
> > import javax.xml.transform.TransformerFactory;
> > import javax.xml.transform.stream.StreamResult;
> > import javax.xml.transform.stream.StreamSource;
> > import java.io.ByteArrayInputStream;
> > import java.io.StringWriter;
> > import java.nio.charset.StandardCharsets;
> >
> > public class Transform {
> >
> >   public static void main(String[] args) {
> >
> >     try {
> >
> >       boolean indent = args != null && args.length == 1 &&
> > "--indent".equals(args[0]);
> >
> >       String template = "<xsl:stylesheet version=\"1.0\" xmlns:xsl=\"
> > http://www.w3.org/1999/XSL/Transform\">\n" +
> >         "  <xsl:template match=\"@*|node()\">\n" +
> >         "    <xsl:copy>\n" +
> >         "      <xsl:apply-templates select=\"@*|node()\"/>\n" +
> >         "    </xsl:copy>\n" +
> >         "  </xsl:template>\n" +
> >         "</xsl:stylesheet>";
> >
> >       String input = "<books>\n" +
> >         "  <book>\n" +
> >         "    <title>test book</title>\n" +
> >         "  </book>\n" +
> >         "</books>";
> >
> >       Templates xsl = TransformerFactory.newInstance()
> >         .newTemplates(new StreamSource(new
> > ByteArrayInputStream(template.getBytes(StandardCharsets.UTF_8))));
> >
> >       final StringWriter result = new StringWriter();
> >
> >       Transformer transformer = xsl.newTransformer();
> >
> >       // as soon as we configure auto-indent the output
> >       transformer.setOutputProperty(OutputKeys.INDENT, indent ? "yes" :
> > "no");
> >
> >       transformer.transform(new StreamSource(new
> > ByteArrayInputStream(input.getBytes(StandardCharsets.UTF_8))), new
> > StreamResult(result));
> >
> >       System.out.println(result.toString());
> >     } catch (Exception e) {
> >       e.printStackTrace();
> >     }
> >
> >   }
> >
> > }
> > -------------------------------
>
>