Convert the plugin pages to Markdown using code blocks

[slachiewicz]

Part of codehaus-plexus/.github#58. All nine plugin pages, following your point that those <li><code> blocks are really code snippets so the inline formatting isn't needed.

You were right, and it reverses the conclusion I reached in #576.

Three commits — please merge or rebase, don't squash, so git log --follow survives the renames.

Why the straight conversion failed

The xdoc marked up method signatures as inline HTML inside a list:

<li><code>public void write( OutputStream output, <i>RootClass</i> root )<br/>
    &#160;&#160;&#160;&#160;throws SAXException, TransformerException</code></li>

Markdown has no syntax for emphasis inside a code span, so doxia-converter emitted `_RootClass_` — underscores rendering literally — and turned each <br/> into a backslash plus a blank line, which closed the list and dumped the rest as plain text.

Treating them as code instead

public void write( OutputStream output, RootClass root )
        throws SAXException, TransformerException

// opens and closes the XML document
public void write( ContentHandler handler, RootClass root )
        throws SAXException

Real monospace, real indentation instead of   runs, and the parenthetical asides become comments on the method they describe.

What is deliberately lost

The <i>/<b> marked which parts of my.model.package.io.sax.ModelNameSAXWriter come from your model and which are literal. A code block can't carry that, so each page now says it in words above the block.

This is a content change, not a faithful format migration, so unlike the other conversion PRs I'm not claiming an identical render.

Three real errors found while reading the sources

1. stax, snakeyaml and jackson all documented their first reader overload with the method name missing — public RootClass ( Reader reader, boolean strict ). It's read, as the xpp3 page shows. Same copy-paste, three pages.
2. The StAX reader delegate was documented in package .io.xpp3. rather than .io.stax. — copied from the xpp3 page.
3. The SAX page's only subsection was headed jackson-writer. I checked all ten pages; that one is isolated.

Dead links replaced

Not a format problem, but I wasn't going to rewrite these pages and leave them:

Was	Status	Now
plexus.codehaus.org/plexus-utils/	DNS gone	codehaus-plexus.github.io/plexus-utils/
plexus.codehaus.org/…/Xpp3Dom.html	DNS gone	plexus-xml apidocs — Xpp3Dom lives there now
code.google.com/p/snakeyaml	Google Code shut down	bitbucket.org/snakeyaml/snakeyaml
wiki.fasterxml.com/JacksonHome	gone	github.com/FasterXML/jackson
dom4j.sourceforge.net/dom4j-1.6.1/	503	dom4j.github.io
Oracle javadoc, Java 1.4.2 and 6	http	Java 8, https

Each replacement checked for a 200.

Verified

All nine build. Titles and authors preserved (including the accented names). Subsection counts match the originals on every page, and so do the method signature counts:

xpp3  7/7   stax 7/7   jackson 6/6   snakeyaml 6/6
dom4j 5/5   jdom 3/3   converters 2/2   sax 4/4

xdoc has no signatures and needed no code block; its features.html link is unchanged from the original.

Supersedes part of #576

#576 says these nine pages stay as xdoc. That carve-out no longer applies — I'll update it. #576 still stands on its own for the Modello overview and the velocity page, including the image map the converter dropped.