21.8a1

Bugfixes & enhancements:

• Fixed the pyperf requirement which was originally pinning pyperf to 2.0.0 by accident.
• Added pre-checks for both pyperf arguments and –format-config configuration so issues are caught earlier.
• The parse task now supports installations of Black as old as 19.3b0!
• Additional metadata is now injected into the pyperf JSON files before being dumped.
• The comments and nested micro targets were tweaked to be more microbenchmark-y.

Project:

• The release process and automation is now ready for use. Which is great since this is the first release planned to be published to PyPI!

21.7.dev2

Bugfixes & enhancements:

• Use posix slashes for target and benchmark names to avoid accidental newlines leading to crashes on Windows
• Added dump command to easily query the source code for any built-in task or target
• Force safety checks to run under the format task to avoid occasional but bad skewing of the data
• Verify that Black is actually available in the current environment before running any benchmarks (and emit a human readable error instead of the original long tracebacky error)
• Arguments can now be passed to the underlying pyperf processes by adding -- followed by such arguments. Useful for debugging stability or slowness issues.
• The command line interface UX has been improved by 1) much prettier help outputs thanks to {pypi}cloup, and 2) official autocompletion support.
• The --targets (or now -t!) option can now also specify a specific target. You can also now repeat the option to additively choose your own custom collection of targets.
• The target name now doesn't include the suffix (eg. black/cache.py -> black/cache). Also the benchmark name format has been changed to $task-name-$target-name (eg. [parse]-[black/cache.py] -> parse-black/cache). Finally format and format-fast were renamed to fmt and fmt-fast respectively. These changes should make the names easier to use (even if less pretty!).
• Redid the targets collection to balance out normal vs micro and also add non-pre-black-formatted code! Run blackbench info for more!
• The output from the info command is now colorful and more informational. Line counts and descriptions are now also emitted.
• The output emitted during the run command is now colourful and a bit more human :)

Project:

• Created initial test suite to make sure blackbench isn't broken on all supported platforms
• Initial documentation including user guide and contributing docs have been written and are hosted on ReadTheDocs.

21.6.dev1

Initial development version of blackbench. Unsurprisingly it is functional but extremely limited. Windows support is actually straight up broken due to a bug :P