Thursday 31st December 2020
Fedora Project
Updated python-invoke to 1.5.0 in Rawhide:
Allow any string-compatible object to be passed to 'Context.cd', enabling use of (for example) 'pathlib.Path' instances (GH#454, GH#577, GH#583, GH#607, GH#681)
Don't silently discard help text for task arguments whose names happen to contain underscores (GH#409, GH#580, GH#611)
Don't silently ignore task help specifiers that don't actually map to the decorated task's arguments (e.g. '@task(help={"foo": "help for foo"})' wrapping a task without a 'foo' argument) (GH#398, GH#580, GH#611)
Allow subcollections to act as the default 'tasks' of their parent collections (via the new 'default' kwarg to '~invoke.collection.Collection.add_collection'); this means that non-trivial task trees can specify, e.g. "use my test subcollection's default task as the global default task" and similar (GH#197)
- Enhanced test coverage in a handful of modules whose coverage was under 90%
'~invoke.context.MockContext' now populates its 'NotImplementedError' exception instances (typically raised when a command is executed that had no pre-prepared result) with the command string that triggered them; this makes it much easier to tell what exactly in a test caused the error
'~invoke.context.MockContext' now accepts a few quality-of-life shortcuts as keys and values in its 'run'/'sudo' arguments:
- Keys may be compiled regular expression objects, as well as strings, and will match any calls whose commands match the regex
Values may be 'True' or 'False' as shorthand for otherwise empty '~invoke.runners.Result' objects with exit codes of '0' or '1' respectively
Values may also be strings, as shorthand for otherwise empty '~invoke.runners.Result' objects with those strings given as the 'stdout' argument
Add a new 'repeat' kwarg to '~invoke.context.MockContext' which, when True (default: False) causes stored results for its methods to be yielded repeatedly instead of consumed (GH#441)
Immutable iterable result values handed to '~invoke.context.MockContext' would yield errors (due to the use of 'pop()'); the offending logic has been retooled to be more iterator-focused and now works for tuples and etc.
- Update the testing documentation a bit: cleaned up existing examples and added new sections for the other updates in the 1.5 release
Automatically populate the 'command' attribute of '~invoke.runners.Result' objects returned by '~invoke.context.MockContext' methods, with the command string triggering that result; previously, users had to do this by hand or otherwise suffered inaccurate result objects (GH#700)
Upgrade '~invoke.context.MockContext' to wrap its methods in 'Mock' objects if the '(unittest.)mock' library is importable; this makes testing Invoke-using codebases even easier