默认情况下是否支持 XmlParser 和 XmlSlurper 命名空间？

Question

当用XmlSlurper 和XmlParser 回答这个与命名空间 (SOAP) 相关的问题 How to read values from XML Request and write into XML Response using Groovy? 时，我意识到我实际上无法判断它们是否非命名空间感知默认情况下。

虽然文档这么说（默认情况下不支持命名空间）：

XmlSlurper

XmlSlurper()

创建一个非验证和非命名空间感知 XmlSlurper 不允许在文档中声明 DOCTYPE。

XmlParser

XmlParser()

创建一个非验证和非命名空间感知 不允许在文档中声明 DOCTYPE 的 XmlParser。

代码正好相反（默认命名空间感知）：

XmlSlurper:

public XmlSlurper() throws ParserConfigurationException, SAXException {
    this(false, true);
}

XmlParser:

public XmlParser() throws ParserConfigurationException, SAXException {
    this(false, true);
}

我还发现这个 answer 声称 XMLSlurper 默认情况下是非命名空间感知的。

Answer 1

我认为文档是错误的，XmlSlurper 和 XmlParser 默认是 namespace-aware：

def text = '''<person xmlns:city="http://localhost/">
    <name>jalopaba</name>
    <city:name>Valladolid</city:name>
</person>'''

// XmlSlurper is namespace aware, since new XmlSlurper() --> new XmlSlurper(false, true)
def slurped_person = new XmlSlurper().parseText(text)
assert 'jalopabaValladolid' == slurped_person.name.text() // :name and :name from city
assert 'jalopaba' == slurped_person.':name'.text() // aware of default namespace
assert '' == slurped_person.'city:name'.text() // city namespace not declared

// but you need to make the resulting GPathResult aware of the concrete namespace
slurped_person = new XmlSlurper().parseText(text).declareNamespace([city: 'http://localhost/'])
assert 'jalopabaValladolid' == slurped_person.name.text() // :name and :name from city
assert 'jalopaba' == slurped_person.':name'.text() // aware of default namespace
assert 'Valladolid' == slurped_person.'city:name'.text() // city namespace declared

// If you make XmlSlurper not namespace aware, it uses the whole element string
slurped_person = new XmlSlurper(false, false).parseText(text)
assert 'jalopaba' == slurped_person.name.text() // just element 'name'
assert '' == slurped_person.':name'.text() // not aware of default namespace
assert 'Valladolid' == slurped_person.'city:name'.text() // using the whole element string (it seems to be namespace aware, but it's not)

// And when it is not namespace aware, nothing changes declaring namespaces
slurped_person = new XmlSlurper(false, false).parseText(text).declareNamespace([city: 'http://localhost/'])
assert 'jalopaba' == slurped_person.name.text() // just element 'name'
assert '' == slurped_person.':name'.text() // not aware of default namespace
assert 'Valladolid' == slurped_person.'city:name'.text() // using the whole element string (it seems to be namespace aware, but it's not)

def c = new Namespace('http://localhost/', 'city')
assert c.name instanceof QName // => new QName('http://localhost/', 'name', 'city')

// XmlParser is namespace aware, since new XmlParser() --> new XmlParser(false, true)
def parsed_person = new XmlParser().parseText(text)
assert parsed_person.toString() == 'person[attributes={}; value=[name[attributes={}; value=[jalopaba]], {http://localhost/}name[attributes={}; value=[Valladolid]]]]'
assert 'jalopaba' == parsed_person.name.text() // default namespace
assert 'jalopaba' == parsed_person[new QName('name')].text() // default namespace
assert 'jalopabaValladolid' == parsed_person[new QName('http://localhost/', 'name')].text() // aware of namespace and empty string as prefix
assert 'Valladolid' == parsed_person[new QName('http://localhost/', 'name', 'it_does_not_matter')].text() // aware of namespace, prefix does not matter
assert '' == parsed_person[new QName('http://wrong_namespace/', 'name', 'city')].text() // aware of namespace, wrong URI with good prefix gets nothing
assert 'Valladolid' == parsed_person[c.name].text()
assert 'Valladolid' == parsed_person.'city:name'.text() // using the whole element string

// With XmlParser not namespace aware...
parsed_person = new XmlParser(false, false).parseText(text)
assert parsed_person.toString() == 'person[attributes={xmlns:city=http://localhost/}; value=[name[attributes={}; value=[jalopaba]], city:name[attributes={}; value=[Valladolid]]]]'
assert 'jalopaba' == parsed_person.name.text()
assert 'jalopaba' == parsed_person[new QName('name')].text()
assert 'jalopaba' == parsed_person[new QName('http://localhost/', 'name')].text() // ignores namespace, only uses 'name' string
assert '' == parsed_person[new QName('http://localhost/', 'name', 'wrong_prefix')].text() // not aware of namespace, the only important thing is prefix
assert 'Valladolid' == parsed_person[new QName('http://wrong_namespace/', 'name', 'city')].text() // wrong namespace but not aware, the only important thing is prefix
assert 'Valladolid' == parsed_person[c.name].text()
assert 'Valladolid' == parsed_person.'city:name'.text()

Answer 2

从 Groovy 2.4.6 开始，文档有误：

/**
 * Creates a non-validating and non-namespace-aware <code>XmlSlurper</code> which does not allow DOCTYPE declarations in documents.
 *
 * @throws ParserConfigurationException if no parser which satisfies the requested configuration can be created.
 * @throws SAXException for SAX errors.
 */
public XmlSlurper() throws ParserConfigurationException, SAXException {
    this(false, true);
}

它实际上创建了一个命名空间感知解析器（第二个参数为 true）。

我还报告了一个错误，即未正确处理命名空间前缀：https://issues.apache.org/jira/browse/GROOVY-7781。幸运的是，另一个用户发布了 PR 来修复。应该有人提交 PR 来解决文档问题，因为这多次给我自己和其他人造成了困惑。