- Python 3.11 compatibility.
- Documentation updates and clarifications.
- Unit test updates.
- Fixed compatibility for Python 3.11 alpha.
- The unit tests were moved out of the
kajiki
package: tests are now under thetests/
directory. - Migrated from Travis CI to GitHub Actions.
- Support dropped for Python 2.7.
- Fixed up lnotab bugs in Python 3.6+ (
ERROR LOOKING UP LINE
messages). - DTD will not be inserted before an
<?xml ...?>
declaration anymore. - Dependency on
nine
dropped. - Dropped integration for TurboGears 1.
- Improved end-to-end testing.
Note
Kajiki 0.9.x will be the last series to support Python 3.4 and 3.5. Users are encouraged to upgrade to Python 3.6+ soon.
- Fixed compatibility for Python 3.10
- Switched the test runner to pytest
- Added the
kajiki
command line interface - Documentation typo corrections
Note
Kajiki 0.8.3 will be the last release to support Python 2.7. Kajiki 0.9 will require Python 3.4+.
- Fixed some HTML entities not being properly recognised
- Fixed unexpected <DOCTYPE> when including templates
- Fixed support for Python 3.8
- Dropped support for Python 2.6 and 3.3
- Improve parsing of
${}
expressions and fix syntax errors reporting in braced expressions. - Improve parsing of documents with comments before the root element
- Allow to replace
gettext
function by providing it in the template context or throughbase_globals
in Loader
- Text for i18n is now extracted ignoring the empty spaces surrounding the text itself. Empty text will always be treated as non translatable nodes for performance reasons.
extract_python
option will now report syntax errors when extracting text for translation.
- Added
extract_python
option to babel message extractor, this allows extracting gettext calls in${}
expressions
- Actually report 0.6 in kajiki/version.py
- Expose
strip_text
option in loader
- Fixed
py:switch
error message wrongly mentioningpy:with
- Support for multiline
${}
expressions - Subsequent text nodes are now squashed into a single text node. This allows translating whole paragraphs instead of single sentences.
- Allow code and function calls inside tag attributes
- Added
strip_text
option to XMLTemplate and i18n collector to ensure leading and trailing spaces are stipped by text nodes (also leads to minified HTML) - Some HTML nodes that do not require being closed but is commonly considered best practice to close them are now emitted with ending tag (IE: <p>)
- Generally improved code documentation to lower entry barrier for contributors
py:attrs
will now emit the attribute name itself or will omit the attribute at all in case ofbool
values for 'checked', 'disabled', 'readonly', 'multiple', 'selected', 'nohref', 'ismap', 'declare' and 'defer',
py:switch
now correctly supports multiplepy:case
statements.- text inside
<script>
and<style>
tags is no longer collected translation. - Syntax errors now report the line and the surrounding code when there is a markup or python syntax error.
- As
py:swtich
discards all its content apart frompy:case
andpy:else
statements it will now correctly report an error when the statements has other content. py:else
will now correctly detect spurious content between itself andpy:if
as the two must be consequential.- Improved code documentation on core classes.
py:with
statement now keeps order of variables, so that variables can depend from each other.- Babel is no longer a dependency unless you want to use the message extractor.
- TranslatableTextNodes are now only generated for non-empty strings
py:with
statement now accepts multiple variables separated by semicolon- Babel message extractor fixed on Python2
- Fix crash on PyPy
- CDATA sections created by the user are now properly preserved
cdata_scripts=False
option inXMLTemplate
allows disabling automatic CDATA for script and style tags.- Autoblocks experimental feature automatically creates blocks from specified tag names.
- Also accept "$." without erroring out. In fact, accept anything.
- Add integration plugin for TurboGears 1
- Accept "$(" without erroring out. Easier to write jQuery stuff.
- There was a showstopper regression in FileLoader. Fixes #1
- Support Python versions 2.6, 2.7, 3.2, 3.3 in a single codebase using the nine library.
- Support HTML entities as well as XML entities in input templates.
- py:include fixed, can see global variables.
- Genshi compatibility: support built-in functions: defined(), value_of() and Markup().
py:def
: Do not crash if a function has no content.py:strip=''
is the same aspy:strip='True'
.- Correctness: escape HTML attribute values.
- Correctness: Always close script tags, even in XML mode.
- Add integration module for the Pyramid web framework.
- Give the FileLoader a path, not just a base directory.
- Documentation improvements, including the need to write CDATA sections.
- Move from Sourceforge to Github.
- Use Travis for continuous integration.
- The whole codebase is formatted according to PEP8.
- Several bugfixes
- Output HTML attributes in alphabetical order (for testability)
- Make Kajiki work on Python 2.4
- Fix Python 2.5 syntax error
- Add support for py:with
- Remove unused babel import that was breaking pip/easy_install
- Python 2.5 fixes
- Correctly strip None attributes and expressions
- Turn off autoescaping in text templates
- Adds i18n support
- Fixes several bugs: [#7], [#9], [#10]