This release drops support for python 3.7, and starts testing under python 3.12.
- Fix typo/braino in
test_buildprog.py::TestIndexBuildProgram::test_produce_artifacts
. (5c01e17)
- We now use ruff for style linting
- Drop support for Lektor 3.2.x
- Test under python 3.11
- Add type annotations
- Convert packaging from setuptools to PDM
- Disuse
pkg_resources
in favor ofimportlib.metadata
- Add pre-commit configuration
- Run black and reorder-python-imports on code
- Drop support for python 2.7 and 3.6.
- Fix deprecation warning from
jinja2
. Jinja2 version 3 is now required.
-
Documentation clarifications, updates and fixes. (PR #2 — Thank you Bart Van Loon!)
-
Add missing requirement
recommonmark
todocs/requirements.txt
. -
Add
docs
enviroment totox.ini
to test that docs will build cleanly.
- Test under python 3.10 and lektor<3.3
No code changes.
Update development status classifier to "stable".
-
Added a
key
field on the index virtual source object. It is an alias to_id
, but is syntactically more self-explanatory. -
The
keys
configuration key has been renamed tokey
. -
When the
key
expression is being evaluted, the record whose key(s) is(are) to be be computed is now available in the jinja context asitem
rather thanthis
.
- Documentation moved from README to Sphinx docs at RTFD.io
-
The
record
argument has been dropped from the (jinja) globalindex_pages
function. (Since indexes can not have multiple parents, it is not necessary.) -
The
parent
configuration key has been renamed toparent_path
. -
The
slug
configuration key has been renamed toslug_format
. -
The
attributes
config section has been renamed tofields
.Though they are not quite like regular Lektor Record fields, they are more field-like than attribute-like. (I.e. access is via getitem rather than getattr.)
Initial release.