Python-clu

Other

* Bump version: 0.3.2 → 0.3.3. [Alexander Böhn]

* Including TOML files in MANIFEST.in. [Alexander Böhn]

Add

* Added new field types and spruced up the existing ones ... also began adding the new configuration schema stuff to the demo “yodogg” project found in tests/. [Alexander Böhn]

Minutiae

* Minutiae. [Alexander Böhn]

Other

* Bump version: 0.3.1 → 0.3.2. [Alexander Böhn]

* Bugfixes in “clu.config” – ... better None-checks in “clu.config.env” ... force-stringify arg in “clu.confiig.settings.Schema.nestify(…)” ... check type before length in “clu.config.fieldtypes.StringField” ... many updates and tweaks to the “clu.fs.appdirectories” module ... testing package EnvBase subclasses and custom schema classes in “test_config.py” ... miscellaneous changes to support all of the above shit. [Alexander Böhn]

* A large miscellany-roundup push, featuring: ... correct-er (i.e. more like the builtin) “update(…)” method’s logic, w/r/t iterable arguments; ... correct-er (i.e. not inconsistent) argument passing/checking in several of the “clu.config.fieldtypes.fields”; ... updates to “clu.config.filebase.FileName” allowing config files to optionally be located and used from “sys.path” directories ... fixes to the aforementioned “clu.config.filebase.FileName” to decouple it from CLU’s app name – config files should all now defer to the “clu.config.filebase.FileBase” respective subclass name information, across the board; ... Lots of docstring tweaks, corrections, and OCD-ish updates; ... The initialization logic of “clu.config.settings.Schema” now takes place in “__new__(…)” rather than “__init__(…)” – this lets subclassers forego calling up to the superclass if they define an “__init__(…)” and clearly separates the concerns; ... Some minor nitpicky updates to the way that the aforementioned “clu.config.settings.Schema” stringification methods work; ... Calling “clu.config.settings.Schema.validate()” returns True if all goes well… this may or may not be gratuitous as it’ll never return False – if validation goes awry at any point the call to “validate()” raises a ValidationError; ... The possible “replaceable endings” in “clu.exporting” are as automatically-generated as possible, and more comprehensive like in general; and finally ... the exporter tests now correctly locate, import, and examine the “yodogg” testing module’s exporter subclasses. [Alexander Böhn]

* WHOOOOPS. [Alexander Böhn]

* Fixing “clu.config.fieldtypes.TupleField” [Alexander Böhn]

* Removed “default” params from NamespacedFieldManager methods. [Alexander Böhn]

* Trimmed a whoooooole lot of dead code. [Alexander Böhn]

* Fixed container field types ... it wasn’t a pretty fix – it involves manually calling a field’s “__set_name__(…)” method – but it works. Blech! [Alexander Böhn]

* Moved file-format-related config stuff into “clu.config.formats” [Alexander Böhn]

* I feel docstring, oh so docstring – [Alexander Böhn]

* Docstrings for “clu.config.settings.Schema” [Alexander Böhn]

* Dead code trim, simplifications, greater helpfulness in exceptions… [Alexander Böhn]

* These tests, they are a-pass-in’ [Alexander Böhn]

* Beginnigs of __repr__ logic for “clu.config.base.{Flat,Nested}” ... also fixes ensuring “clu.dicts.merge_two(…)” will actually work. [Alexander Böhn]

* More wide-ranging updates to the “clu.config” machinery ... Added a “clone()” method! ... Fixed “clu.dicts.merge_two(…)” to be properly generic, and thus support these new mapping types! ... cleaned the dead code out of “clu.config.tomlfile”! ... AND MORE!!!! [Alexander Böhn]

* Had to rename “clu.config.toml” as “clu.config.tomlfile” ... to avoid confusing `import` – like it was somehow doing py2-ish implict relative imports, “import toml” from within the file that defined “clu.config.toml” would yield “clu.config.toml” – what the fuuuuuuuuuuuck? So I changed it. [Alexander Böhn]

* Many updates to “clu.config” … ... docstrings abound! ... fixes for the “clu.config.base.Flat.{keys,values}(…)” methods! ... “Flat.nestify(…)” and “Nested.flatten(…)” allow reformatting namespaced-dictionary data losslessly between internal formats! ... Expanded the “find_file(…)” method and moved it and its many helper class-methods from “clu.config.filebase.FileBase” up to “clu.config.filebase.FileName” – the key difference being that the latter is a mixin class and can be reused outside of the FileBase implementation(s)! ... Most of the intermediate file-finding-related methods can take either additional or overriding parameters which will be great for testing!! ... other miscellaneous shit I am sure!!! [Alexander Böhn]

* Self-test introspection for the “clu.config.toml” module. [Alexander Böhn]

* Getting ever-closer – adding “clu.config.toml” [Alexander Böhn]

* Continuing apace on “clu.config.filebase” [Alexander Böhn]

* Fleshing out “clu.config.base”; staring on “clu.config.filebase” [Alexander Böhn]

* Start of the “clu.config” module. [Alexander Böhn]

* Moved a few enum-alias-related predicates into “clu.typology” [Alexander Böhn]

* First stab at slash-operators for Directory filesystem types. [Alexander Böhn]

* Trying to get the PyPy testenv to actually use PyPy, like it used to. [Alexander Böhn]

Add

* Adding “clu.constants.enums” to the hardcoded module list. [Alexander Böhn]

* Adding “clu.exporting.Exporter” to “clu.repl.columnize” [Alexander Böhn]

Minutiae

* Minutiae in the “show-modules.py” script. [Alexander Böhn]

Other

* Bump version: 0.3.0 → 0.3.1. [Alexander Böhn]

* Some tox.ini adjustments. [Alexander Böhn]

* Raising when calling “flatten(…)” on a nonexistant Directory. [Alexander Böhn]

* Confirming the new script directory location in the testsuite. [Alexander Böhn]

* Tweaks to the new display capacities of the “show-modules.py” script. [Alexander Böhn]

* Removed unused import. [Alexander Böhn]

* The “show-modules.py” script actually shows all the modules. [Alexander Böhn]

* Vendoring in a “columnize” implementation ... courtesy https://github.com/rocky/pycolumnize. [Alexander Böhn]

* Updating the hardcoded module list. [Alexander Böhn]

Other

* Bump version: 0.2.10 → 0.3.0. [Alexander Böhn]

* Test for “clu.fs.filesystem.script_path(…)” no longer xfails. [Alexander Böhn]

Add

* Added a stub clu/__main__.py file (all it does now is print the version) [Alexander Böhn]

Other

* Bump version: 0.2.9 → 0.2.10. [Alexander Böhn]

* Moved “scripts” directory into the CLU module base proper. [Alexander Böhn]

* F-strings in “clu/__init__.py. [Alexander Böhn]

Add

* Added a 'clu-boilerplate' console script entry point ... which echoes out the (mercifully very short) boilerplate you need to use CLU in a new Python module. [Alexander Böhn]

* Added the Exporter stuff to “clu.testing.utils” [Alexander Böhn]

* Added a “temporaryname” fixture-factory function to “clu.testing” [Alexander Böhn]

Minutiae

* Minutiae. [Alexander Böhn]

Other

* Bump version: 0.2.8 → 0.2.9. [Alexander Böhn]

* Stubbing out migration methods in “clu.keyvalue.CLUInterface” [Alexander Böhn]

* Enabled versioning by default in “clu.keyvalue.interface” [Alexander Böhn]

* Installing CSV-related lambdas from an earlier CLU iteration. [Alexander Böhn]

* Commented out hardcoded length-check lines in exporter tests. [Alexander Böhn]

* “weakref.WeakValueDictionary” is the new “collections.OrderedDict” ... as far as the “clu.fs.filesystem.TypeLocker” internal registry is concerned at least doggie. [Alexander Böhn]

* Revised the “clu.fs.filesystem.TypeLocker” registry metaclass ... got rid of the “__prepare__(…)” method that just returned an `OrderedDict` – it’s 2019 and our Python 3.7 dicts are more ordered than `OrderedDict` (and also less janky in the repr). ... Properly assign “__name__” and “__qualname__” – and also our own “__lambda_name__” – to each lazy-static “directory(…)” method that TypeLocker creates, using the same semantics and logic as “clu.exporting.ExporterBase.export(…)” ... Updated docstrings and nota-benne comments therein ... Also threw in some filesystem-module tests updates and other assorted miscellany. [Alexander Böhn]

* Renamed “suffix(…)” to “re_suffix(…)” to clarify things. [Alexander Böhn]

* Assorted touchups in “clu.exporting” ... to wit: a) made ValueDescriptor a proper data descriptor (if read-only) b) made both “__class_getitem__(…)” calls throw the same sorts of exceptions if they are passed the wrong shit c) added a proxied “clu.exporting.ExporterBase.items()” method d) something else nifty that I can’t remember just now, oh well. [Alexander Böhn]

* Reshuffle ... WHOOOOOOOOOOOPS. [Alexander Böhn]

* Including “.pyc” files in package index suffixes. [Alexander Böhn]

* Module-file suffixes for “path_to_dotpath(…)” come from “importlib” [Alexander Böhn]

* Stacking exports atop lru_cache decorators in “clu.fs.misc” [Alexander Böhn]

* Lowercasing all suffixes before comparison. [Alexander Böhn]

* Limiting function cache size on regex match/search functions. [Alexander Böhn]

* Caching “clu.fs.misc.re_{matcher,searcher}” functions ... using default-value “functools.lru_cache(…)” RN. [Alexander Böhn]

* Nitpick on the variable name inside the returned lambda ... within “clu.fs.misc.re_{matcher,searcher}” – it has, up to this point, been `searching_for`, which is wrong as this variable contains the string that itself is being searched; truncating the name to `searching` satisfies my OCD in this case (read it and you’ll see if you find this summary confusing, tensewise). [Alexander Böhn]

* Un-exporting two module constants in “clu.repl.banners” [Alexander Böhn]

* Un-exporting two module constants in “clu.enums” [Alexander Böhn]

* Un-exporting two module constants in “clu.fs.filesystem” ... get that comparison failure rate down doggie. [Alexander Böhn]

* Generalizing “clu.fs.misc.suffix_searcher(…)” [Alexander Böhn]

* Using “clu.fs.misc.gethomedir()” in “AppDirs” tests. [Alexander Böhn]

* Using new export “clu.fs.misc.gethomedir()” in the filesystem code. [Alexander Böhn]

* Nota benne. [Alexander Böhn]

* Moved one-off “gethomedir()” lambda to “clu.fs.misc” ... and exported it from there with a docstring and everything. [Alexander Böhn]

* Refactored the “clu.fs.filesystem.rm_rf(…)” test with “countfiles(…)” [Alexander Böhn]

* Passing through “parent” in the “temporaryname” fixture factory. [Alexander Böhn]

* Error-checking in “clu.fs.filesystem.TemporaryName.filesize(…)” [Alexander Böhn]

* Test fixture manscaping. [Alexander Böhn]

* One more “pth” → “path” changeover. [Alexander Böhn]

* Starting to normalize the argument names in “clu.fs.filesystem” ... tests still pass, so far…¬ [Alexander Böhn]

* Prevent leakage from the “clu.fs.filesystem.TypeLocker” metalclass ... specifically, all classes for which TypeLocker was their meta – we still need a good word for that – would receive a “types” attribute that was a reference to an OrderedDict full of all of those classes, as kept internally by TypeLocker for its own housekeeping purposes. That was a downside of its use, as this attribute was kind of hanging out in the open, using a fairly common name with no underscore-prefixing (something I kind of loathe, personally, but that’s me) or other indication of what it was or what it was for or how shit could break if it were to be improperly fucked with. ... This solves the problem by overshadowing the “types” attribute with a read-only “clu.exporting.ValueDescriptor” instance on all generated classes. [Alexander Böhn]