[MSHARED-258] PrettyPrintXmlWriter encoding of \u0000 in xpp3dom attribute incorrect/different from p-u xpp3dom

[jira-importer]

Kristian Rosenvold opened MSHARED-258 and commented

When porting surefire to m-s-u I came across the case where
unicode \u0000 gets encoded as � in an xml attribute in the prettyprintxmlwriter. This is probably the reason why the PrettyPrintXmlWriter was forked into surefire originally.

Now from SUREFIRE-456 it seems like it's specification-wise illegal to do this encoding, but it does actually preserve the character value of non-printable characters.

So the more I type on this issue, the more it seems like the forked PPXW actually does the best-effort "right" thing when it comes to xml encoding "any" string as long as we want to stay human readable....?

I need some input on this one ;)

---

Affects: maven-shared-utils-0.1

[jira-importer]

Kristian Rosenvold commented

OK, this turned out to be a can of worms, since there is no legal way to represent the 0 character in xml.

Since the xml file tries to stay compatible with ant, this is not really going to change. There are basically only 2 options:

Encode illegally (which is the current solution, giving � or just strip it away).

I think this issue is a won't fix, I'll let it simmer for a day or two before giving up ;)

[jira-importer]

Joerg Schaible commented

Facts in short: Null character encoding is always illegal in XML. Control characters encodings are illegal in XML 1.0, but valid in XML 1.1 (simply write appropriate XML header).

Long story: Jason used ages ago the code for Xpp3Dom in XStream and took the PrettyPrintWriter also in Plexus. The code evolved in XStream also over the years though ;-)

See special note and linked XML specs in http://xstream.codehaus.org/javadoc/com/thoughtworks/xstream/io/xml/PrettyPrintWriter.html

[jira-importer]

Kristian Rosenvold commented

It would seem like maybe using xml 1.1 and stripping 0 would be an option that'd stay reasonably legal. There is really no use-case for preservin g the 0 as far as I know

[jira-importer]

Michael Osipov commented

This issue has been auto closed because it has been inactive for a long period of time. If you think this issue still persists, retest your problem with the most recent version of Maven and the affected component, reopen and post your results.