Age | Commit message (Collapse) | Author |
|
Encoding quotes inside element contents may be not necessary
|
|
that method.
|
|
Change to comments does not seem to produce different output under
any cases. See Issue #19.
|
|
|
|
classes a little easier.
|
|
level was a wrapper for parsing. Moved the serializing interface to be in the same form as the parser for consistency.
|
|
change to commonize how we override classes to be more consistent.
|
|
This caused a trailing whitespace after the PI contents and before
the PI closing tag.
|
|
Thanks to @MrElectronic for catching this one.
|
|
|
|
a fragment.
|
|
|
|
encounter this class once. It's a container like DOMDocument.
|
|
|
|
|
|
attributes with a value of 0.
|
|
|
|
|
|
|
|
and MathML output.
|
|
default option setup and have it still make sense.
|
|
|
|
|
|
|
|
|
|
|
|
wholeText on a text node holds the text of logically adjacent text. That means a new line followed by cdata following by spaces and/or a new line will all be in the cdata wholeText. This change fixes our handling to use data rather than wholeText.
Also fixed the cdata regex tests.
|
|
The Traverser now simply walks through a document. The OutputRules convert
the nodes into output html. The rules is a configurable options. By default
OutputRules will generate html close to the html5 that was parsed. Alternate
rule implementation (e.g., minify rules, pretty spacing rules) can be set
as the default or on an individual case.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
then the regular expressions and twice as fast as the next fastest method I tested.
|
|
|
|
|
|
for it.
|
|
|