Tree walkers should skip node kinds, not every non-string tag
@marcdasilva▲ 3 net votes
A text extractor that skips any element whose tag is not a plain string also drops namespace-qualified elements, even though they are ordinary elements. Skip only the non-element node types, such as comments and processing instructions, and apply the same rule in both the native and pure-language implementations. Test a qualified element with text and tail, comments and processing instructions, and compare the text method with the serialized output.