Kid Version History
===================


2005-03-14 0.6.3
-----------------

  * Fixed issue introduced with py:extends that broke kid on Python 2.2.
    See: <http://www.kid-templating.org/trac/ticket/50>

  * Fixed various issues with HTML output methods including incorrect
    DOCTYPEs, attribute casing, etc.
    See: <http://www.kid-templating.org/trac/ticket/49>

  * Applied Tim Gerla's patch to add a ``--strip-dest-dir`` option that acts
    like the argument with same name in Python proper. This is mainly useful
    for package maintainers.
    See: <http://www.kid-templating.org/trac/ticket/52>


2005-03-09 0.6.2
-----------------

  * Fixed bug with short-form elements (<br />) not being output properly.
    Long form was always being used.

  * Added support for custom serialization routines including XHTML and HTML.
    See: <http://www.kid-templating.org/trac/ticket/4>

  * Fixed bug where ElementTree objects passed into kid were being modified
    (children removed) during template execution.
    See: <http://www.kid-templating.org/trac/ticket/44>

  * This file is now included the source distribution again.
    See: <http://www.kid-templating.org/trac/ticket/43>

  * Fixed various unicode related issues (mainly with the XML() function).
    See: <http://www.kid-templating.org/trac/ticket/45>


2005-03-06 0.6.1
-----------------

  * Fixed bug (again) with comments and PIs not making it through to output.

  * Fixed bug with text content being dropped when passed to the XML()
    function.

  * No longer escaping ( > | ' | " ) in output (except in attributes where
    quotes are escaped.

  * Fixed bug with doc not shipping with css files.

  * Namespace prefix generation and handling should be working well at this
    point.


2005-03-05 0.6
---------------

For detailed information see: <http://www.kid-templating.org/doc/kid/0.6/notes.html>

  * Namespace URI changed from "http://naeblis.cx/ns/kid#" to
    "http://purl.org/kid/ns#".

  * py:omit is now py:strip.

  * Expression substitution changed from {} to ${} and $name.

  * Python interface is more like Cheetah.

  * Support for cElementTree.

  * Template inheritence (py:extends).

  * Match Templates (py:match)

  * Comment wart is no longer.

  * ElementTree objects can be referenced directly in content constructs.

  * Major updates to documentation.


2005-02-14 0.5.2
-----------------

  * Fixed bug with namespace prefix being stripped in xml:* attributes (such
    as xml:lang). This could sometimes result in ill-formed XML.

  * Examples have been brought up to date.

  * Added functionality to kid command line script. (Christoph Zwerschke)
    <http://www.kid-templating.org/trac/ticket/17>


2005-02-11 0.5.1
-----------------

  * Fixed problem with brace interpolation sometimes not working in text
    content due to pulltree reporting text in multiple hunks.

  * Should be able to build RPMs with `python setup.py bdist_rpm` now.

  * Comments and Processing Instructions should now be passed through
    properly. These were previously being omitted from output. There may be
    an argument in the future that turns comment output off.

  * Template modules used to import various names from ElementTree and
    pulltree into the module namespace. This can cause problems if a
    template imports other stuff with the same name. The Element and
    ElementTree imports have been changed to _Element and _ElementTree to
    avoid clashing. (Christoph Zwerschke)

  * Add support for specifying a character encoding when loading kid
    templates. (Christoph Zwerschke)

  * Fixed bug in kid:attrs where python reserved words could not be used as
    keyword arguments. (Christoph Zwerschke)


2005-01-26 0.5
---------------

  * Changed `<?kid?>` processing instruction name to `<?python?>`. Not sure
    why I didn't think of this in the beginning. Note that `<?kid?>` is still
    supported but deprecated. I may leave it in forever just in case there's
    ever a clash with another `<?python?>` PI.

  * Changed stuff everywhere to use `py:` as namespace prefix instead of
    `kid:`.

  * Applied `py:attrs` patch from Christoph Zwerschke. More detail on this
    in the following thread:

    <http://article.gmane.org/gmane.comp.python.kid.general/11>

    The concept is identical to `tal:attributes` with minor syntax
    differences. For example:

        <elem x="10" y="20" py:attrs="y=30, z='foo'"/>

    Will result in the following output:

        <elem x="10" y="30" z="foo" />

  * Allow brace interpolation in normal text content. This isn't set in stone
    but there's been enough requests for it and I've found myself wanting
    something less verbose than `kid:content`.

        <test>the current time is: {time.strftime('%C %c')}</test>

    Yields something like this:

        <test>the current time is: 20 Wed Jan 26 01:49:56 2005</test>

    This can serve as a terse alternative for clumsy situations like the
    following:

        <test>
          the curent time is: <span py:content="time.strftime('%C %c')"/>
        </test>


2005-01-18 0.4.2
-----------------

  * Fixed `.pyc` file generation on import for real this time (thanks:
    Christoph Zwerschke).

  * Fixed problem with using non-string values in attribute value templates
    "{}". For instance, `<elem attr="{1234}"/>` would fail with a type error.
    (thanks: Dagur Páll Ammendrup)

  * Added support for DOCTYPE declarations. They were being chopped previously
    making it impossible to generate validating.. well.. anything..
    (thanks: Dagur Páll Ammendrup)

  * Added support for omitting namespace declarations used in the source
    template. A template can set the `omit_namespaces` variable to a list of
    namespace URIs that should be omitted.

    The following example results in the `xmlns:kid` and `xmlns:x`
    declarations being dropped:

        <?xml version="1.0" encoding="utf-8"?>
        <?kid #
        omit_namespaces += ['urn:x']
        ?>
        <html xmlns:kid="http://purl.org/kid/ns#"
              xmlns:x="urn:x">
          ...
        </html>

    The `omit_namespaces` variable is set to `['http://purl.org/kid/ns#']`
    by default, hence the weird appending of the additional namespaces to
    exclude to the list.

    (thanks: Christoph Zwerschke)

  * Setting the environment variable KID\_OUTPUT\_PY to something True will
    dump out `.py` source files along side of `.pyc` files when templates
    are imported.

  * Make `test_kid.py` portable for upcoming Windows testing. This mostly
    consisted of getting rid of hard coded path separators.

  * Added a few unicode tests to make sure there isn't any funny business with
    unicode in templates.


2005-01-12 0.4.1
-----------------

  * Fixed bad output when running templates with the `kid` command.
    There was a print statment in there that shouldn't have been and the
    template was being executed twice. (thanks: Ross Burton)

  * Fixed compiler not outputting `.pyc` files correctly. (thanks: Ross Burton)

  * Updated tutorial documentation.


2005-01-10 0.4
---------------

  * Initial crack at reusable templates (damn if that word doesn't mean five
    things in this system. are you guys confused yet?). See the kid language
    spec on the `kid:def` attribute. It provides the equivalent of a Python
    function.

  * Templates now extend the base class: `kid.BaseTemplate`. When you import a
    kid template module, a class named `Template` is available from the
    module.

    You can pass keyword arguments to the Template class constructor and have
    them be accessible to all template expansion performed on the
    instance. The template code accesses these variables by using `this`.

  * The following methods are available on Template instances and template
    modules:

    * `serialize(encoding='utf-8', **kw)` - Execute the template and return
      a string.

    * `generate(encoding='utf-8', **kw)` - Like `serialize()` but returns a
      generator that continuously yields a string until the template is
      finished.

    * `write(file, encoding='utf-8', **kw)` - Execute the template and write
      output to the specified filename or file like object.

    * `pull(**kw)` - Generator that yields pulltree events for each infoset
      item generated by the template.

  * Template methods now take keyword args. Any arguments passed are available
    by name in the template as local variables.

  * Added support for escaping curly braces in attributes. Works like XSLT:
    '{{' and '}}'. Note also that you don't _need_ to escape braces unless you
    have both ends: {example}. If you just have a single open or a single close
    brace, you shouldn't need to escape anything. It won't hurt anything if you
    do however.


2004-12-21 0.3
---------------

A couple quirks have been straightened out since 0.2 but the primary reason I'm
putting out a new release now is because of a license change. Kid is now
licensed under an MIT style license. All versions prior to 0.3 are licensed
under the GPL. This is due to a couple of reasons:

  * I stole a bit of code from Quixote, which is licensed under a GPL
    incompatible license.

  * I plan on using/distributing various GPL and non-GPL licensed libraries
    with systems that use Kid.

Quite a bit else has shifted underneath the hood during this release. I tried
to implement XSLT-like templating but found that it was going to be extremely
messy to implement on top of SAX (if possible at all). That combined with an
absolute refusal on the part of the stingy author to bring in `xml.dom` in any
way led to [effbot's ElementTree][et], which is completely in line with Kid's
goals around simplicity. This opened a few doors for new features in this
release but mostly just took up a lot of time.

Here's the full list of major modifications since 0.2:

  * Rewrite parser to be [ElementTree][et] based instead of SAX based. This is
    in preparation for templating. Note that this means Kid now requires
    ElementTree where before it used only modules from the standard library.

  * `document()` function for bringing in external XML content from a file or
    URL. Acts *kind of* like the XSLT's `document` XPath extension function
    sans relative URI resolution and the ability to resolve multiple resources.

  * `XML()` function for bringing in XML content that is a string. Basic usage
    is to wrap XML() around a well-formed XML string returned from a python
    module function call.

  * New pulltree module provides a streaming pull interface to ElementTree.
    [Check it out!][pulltree] `pulltree.py` is not dependent on Kid at all. I
    may split it out into a separate project if there's interest.

  * Support for `kid:replace`. (I thought I added this in 0.1)

  * Tests are now actually validating that output is matching expected
    output.

  * Tests pass on Python 2.2.

  * `kid:repeat` is now `kid:for`. Templates that used `kid:repeat` will break
    under this version.

  * Starting with 0.3, Kid will be licensed under the "MIT License" as
    specified [here](http://www.opensource.org/licenses/mit-license.php). This
    is due to license compatibility issues with the ElementTree package,
    compatibility issues with a bit of code taken from Quixote in
    `kid/importer.py`, and to ensure we have the ability to distribute Kid with
    applications written atop Quixote.

    <small>
    Note: the GPL isn't viral - everything else is viral :(
    </small>

[pulltree]: http://cvs.sourceforge.net/viewcvs.py/splice/kid/pulltree.py?view=markup
            "pulltree.py"
[et]:       http://effbot.org/zone/element-index.htm
            "ElementTree Overview"


2004-12-01 0.2
---------------

Initial public release.

  * Added documentation.

  * Makefile is halfway useful.

  * Broke up kid.py into three separate modules and placed in a
    kid package.

  * distutils install.

  * kidc command for compiling templates.

  * kid command for running templates from the command line.

  * Added basic and cgi examples under `examples` directory.


2004-11-29 0.1
---------------

Barely useful. No documentation.

  * Supports the following attributes: kid:content, kid:if, kid:repeat,
    kid:omit.

  * `<?kid?>` processing instruction for embedding code blocks.