Textpattern Support Forum

A completely rewritten parser just landed in the 4.0 development tree

• Full tag nesting support (unlike the current 4.0.6 parser), so this works:

<txp:if_section="something">
  do this
<txp:else />
  <txp:if_section="different">
     do that
  <txp:else />
     and otherwise do this
  </txp:if_section>
</txp:if_section>

• Attribute value escaping: only the delimiter characters (single or double quotes) must be escaped. All other characters need no escaping inside attribute values. Examples:

<txp:tag attribute="'single' ""double"" quoted" />
<txp:tag attribute='''single'' "double" quoted' />
<txp:tag attribute="< > ' &" />

• Unquoted attributes are deprecated and will give a warning in testing/debug mode:

Wrong: <txp:tag attribute=value />

• Attribute parsing for attribute values enclosed in single quotes (better know as ‘wondertag’ support). Provided you use the proper quote escaping, this works till unlimited depth.

<txp:tag not_parsed="value" parsed='<txp:sometag attr="value" />' />

The new core parsing function (parse) is 50-100% slower than the old parser, but since 4.0.6, several other parsing related functions have been optimized for speed. I did a few tests with a fresh TXP install. Compared to 4.0.6, runtime (minus query time) was actually 5% faster! The actual speed difference will probably vary depending on how complex (tag nesting level) your templates are, but I don’t expect dramatic speed differences with 4.0.6.

Since this is a rather big change for what we’d like to think of as a stable TXP version, it could really benefit from some testing.

Last edited by ruud (2008-04-19 18:17:07)