Python-clu

Add

* Added a “consts” fixture to the pytest plugin. [Alexander Böhn]

Other

* Bump version: 0.4.10 → 0.5.0. [Alexander Böhn]

* Made “clu.importint.ProxyModule” a real thing ... with tests (both inline and unitary), docstrings and notes, differentiated support functions and classes… YOU NAME IT ... !!!!!!!!!!!!!!!!!! FUCK YES !!!!!!!!!!!!!!!!!!! [Alexander Böhn]

* Further notations, errata, and error-message minutiae. [Alexander Böhn]

* Bespoke (aka hack-tastic) one-off proxy-module typename reporting. [Alexander Böhn]

* De-redundifying proxy-module “__getattr__(…)” error handling. [Alexander Böhn]

* Position one arg for module name, keyword-only arg for docstring. [Alexander Böhn]

* Docstring for experimental proxy-module “__init__(…)” call. [Alexander Böhn]

* Uncluttering proxy module namespace via function inlining. [Alexander Böhn]

* Trimming class-module instance methods in “__execute__()” [Alexander Böhn]

* Trimming intermediate target lists from proxy module internals. [Alexander Böhn]

* Threw in an additional assert in old test, just to be safe. [Alexander Böhn]

* Updated “clu.typology” assertion regarding “clu.extending” ancestry. [Alexander Böhn]

* Experimental class-based module-wrapper proxy thing in tests ... the meat of this currently resides in “test_importing.py”, q.v. test function ‘test_module_dict_proxy_idea’ supra. for the module-class proxy reference implementation, a working example subclass and example code that does not fail to run ... basically it’s a ChainMap for module attributes (like literally as this implementation uses “clu.dicts.ChainMap” internally) ... also internally, the sub-sub-sub-sub-metaclass base for all class-based modules inherits from “clu.extending.Extensible” – which that type itself inherits from “clu.abstract.NonSlotted” instead of plain ol’ ‘type’ ... the implementation classes for the “clu.extending.doubledutch” decorator – “DoubleDutchRegistry” and “DoubleDutchFunction” – now leverage a number of CLU- and standard-library-based ABCs. [Alexander Böhn]

* The “clumods” pytest fixture now depends on the “consts” fixture. [Alexander Böhn]

* Session-scoping the “greektext” pytest fixture. [Alexander Böhn]

* Moreso deployment of the new “consts” pytest fixture. [Alexander Böhn]

* Updated comment note on module inclusion criterion. [Alexander Böhn]

* Import shuffle. [Alexander Böhn]

* Killed trailing whitespace. [Alexander Böhn]

* Generalized and centralized import-all-modules logic ... created a new module “clu.all”, containing two functions: • “import_all_modules(basepath, appname)” → Imports all modules, both file-based and class-based, from the app «appname» within the package rooted at «basepath» • “import_clu_modules()” → Imports all CLU-specific modules – a convenience call for “import_all_modules(consts.BASEPATH, consts.PROJECT_NAME)” ... the “clu.testing.pytest.clumods” fixture now simply delegates to a call to “clu.all.import_clu_modules()” ... the clu-module-importing function in ye olde “show-modules.py” script has been replaced with “clu.all.import_all_modules(…)” ... no specific tests have been added because this shit is already super-100%-plus covered by existing test code, doggie. [Alexander Böhn]

* Re-instating “rm_rf(…)” usage in pytest plugin’s exit handle. [Alexander Böhn]

* Using “shutil.rmtree(…)” in “clu.fs.filesystem.rm_rf(…)” ... instead of all of my tortured bespoke logic that apparently did not quite work right. [Alexander Böhn]

* Noting possible one-liner for “installed_appnames()” impl. [Alexander Böhn]

Other

* Bump version: 0.4.9 → 0.4.10. [Alexander Böhn]

* Updated/refactored some of “clu.fs.pypath” ... “pypath.append_path(…)” has been renamed “pypath.add_path(…)”, and it now accepts a keyword-only argument ‘prepend=True’ to, y’know, prepend its payload to ‘sys.path’ instead of appending. ... “pypath.remove_invalid_paths()” calls ‘site.removeduppaths()’ before doing anything to ‘sys.path’ ... There’s a new convenience function “pypath.enhance(…)” which is basically sugar for calling “remove_invalid_paths()” ahead of calling “add_path(…)” – which as already noted now also includes a call to ‘site.removeduppaths()’ ... the REPL script imports “clu.fs.pypath” as a module, instead of picking through its exported functions ... many tests make use of new “clu.fs.pypath.enhance(…)” function. [Alexander Böhn]

* Moved the “pytester” requirement into the CLU pytest plugin proper. [Alexander Böhn]

* Testing and pytest support for “clu.dispatch” ... new “clu.constants.consts” item ‘USER’, value of the current users’ username ... rework of “clu.fs.filesystem.rm_rf(…)” logic ... The “clu.testing.pytest” plugin now implements a pytest hook function “pytest_sessionfinish(…)”, which in turn conditionally binds an exit handler – using “clu.dispatch.exithandle” – that deletes any stray pytest temporary-file artifacts left over upon interpreter shutdown … namely, anything in the directory $TMPDIR/pytest-of-$USER – which stubbornly would not remove itself and (according to the policy of pytest’s code for this, apparently) just keeps accumulating piles of cruft every time ‘pytest’ was executed ... All in aid, really, of the one new test, in “test_dispatch.py”, which makes use of the “pytester” built-in pytest plugin to cleanly test exit handlers; see the source of same for details. [Alexander Böhn]

* Updated the bpython REPL script for the ‘dispatch’ update. [Alexander Böhn]

* Moved “clu.shelving.dispatch” down to “clu.dispatch” ... as it is clearly bigger than just the nascent ‘shelving’ module. [Alexander Böhn]

* Made “clu.fs.filesystem.TemporaryFileWrapper” an explicit Iterable ... as in, it inherits from ‘collections.abc.Iterable’ ... also added 'pytester' to the test plugins loaded in conftest.py. [Alexander Böhn]

Add

* Added ‘has_appname’ to “clu.exporting.ExporterBase.__dir__(…)”’s filter. [Alexander Böhn]

Other

* Bump version: 0.4.8 → 0.4.9. [Alexander Böhn]

* Split off async parts of “clu.abstract.ManagedContext” ... into “clu.abstract.AsyncManagedContext” [duh] ... also added tests for the former. [Alexander Böhn]

* Exclude async methods on lower pythons from “clu.abstract.ManagedContext” [Alexander Böhn]

* Re-enabling test for qualified naming of constants. [Alexander Böhn]

* Elected to use context-managed exporting in “clu.exporting” itself. [Alexander Böhn]

* Trimmed and tweaked “clu.exporting” generally – ... moved “clu.exporting.rename” to “clu.naming” where arguably it has always belonged ... removed nearly all method-level imports in “clu.exporting” in favor of module-level; so far, so good ... made a couple minor tweaks to “clu.exporting.ExporterBase”, of which the most notable is that “clu.exporting.ExporterBase” is now a context manager; doing: [Alexander Böhn]

exporter = Exporter(path=__file__)

with exporter as export:

export
def yodogg():
...

… now works, and makes a certain amount of sense

* Removed old “Python 3”-specific REPL module. [Alexander Böhn]

* Removed old Bash REPL stub. [Alexander Böhn]

* Fixed docstring *again* pt. II. [Alexander Böhn]

* Fixed docstring *again* [Alexander Böhn]

* Use current if “importables(…)” called with Falsy subdirectory. [Alexander Böhn]

* Docstring minutiae. [Alexander Böhn]

* REPL script updates. [Alexander Böhn]

Other

* Bump version: 0.4.7 → 0.4.8. [Alexander Böhn]

* Typographic eratta en extremis. [Alexander Böhn]

Other

* Bump version: 0.4.6 → 0.4.7. [Alexander Böhn]

Add

* Added a “clu.shelving.dispatch.shutdown(…)” function ... like “clu.shelving.dispatch.trigger(…)” but with an actual call to ‘sys.exit(¬)’ at the end ... also more bells & whistles to “clu.shelving.redat.RedisConf” have been grafted on, somehow. [Alexander Böhn]

* Added a bunch of async shit I don’t quite understand. [Alexander Böhn]

Other

* Bump version: 0.4.5 → 0.4.6. [Alexander Böhn]

* SIG-WINCH!!!!! [Alexander Böhn]

* Logging format config manscaping. [Alexander Böhn]

* Trimmed dead code. [Alexander Böhn]

* Tweaking shutdown logic. [Alexander Böhn]

* Minor tweak to zipfile artifact save logic. [Alexander Böhn]

* Even more “clu.shelving.dispatch” minutiae. [Alexander Böhn]

* Exit handle functions execute properly from signal handlers. [Alexander Böhn]

* More tweaks to async signal-handler demo code. [Alexander Böhn]