jhannes / eaxy Goto Github PK

Event driven is annoying. One alternative is to use a lazy Iterable. The syntax could be something like this:

for (Element element : Xml.iterate(someReader).find("foo", "bar", "baz")) {
}

The element contains a full subtree but eaxy would discard it when Iterator.next() is called.

This feature is of course most useful for documents that could be thought of as consisting of several sub-documents.

first, firstOrNull, single, singleOrNull to deal with the four possible restrictions (accept many or not, accept no hits or not).

Convenience method for finding exactly one element by path: Finding a set of
elements, insuring that there was exactly one result, and then
grabbing the first element in the set.

                    ElementSet verticesSet = mesh.find("vertices");
                    if (1 != verticesSet.size()) {
                        throw new RuntimeException("Found " +

verticesSet.size() + " vertices sets for geometry id=" + geometry.id()

• " name=" + geometry.name());
  }
  Element vertices = verticesSet.first();

Outstanding question: Would you want them to be searching arbitrary depth or direct path?

When it comes to finding by attribute value this is supported with the syntax find("verticies[class=rounded]"), or find(ElementFilters.attr("class", "rounded")).

Provide chained or fluent-style element query filters:

Example:

element.query().tagName("verticies").attr("class", "rounded").list()

Running org.eaxy.usage.ValidationTest
Tests run: 2, Failures: 0, Errors: 1, Skipped: 0, Time elapsed: 0.189 sec <<< FAILURE!

Tests in error:
shouldAcceptValidXml(org.eaxy.usage.ValidationTest): cvc-complex-type.3.2.2: Attribute 'role' is not allowed to appear in element 'msg:recipient'.

We love this easy-to-use xml library and use it more and more in stead of e.g. JAXB.
There are however, some utility methods which are hidden from outside use with no apparent reason. In specific, we'd like to be able to access DomTransformer.createElement, but there are others, e.g. org.eaxy.DomTransformer.createName, IOUtil.slurp(BufferedReader reader) etc.
Regards,
Erik Dybdahl

add a method Document.setEncoding(encoding) not hardcode UTF-8

and() element filter should not require filter name in constructor -- default to "&" perhaps

return a ordered map of all attributes on this element.