Every one of the projects in this repository is available at the canonical
URL git://git.madduck.net/madduck/pub/<projectpath> — see
each project's metadata for the exact URL.
All patches and comments are welcome. Please squash your changes to logical
commits before using git-format-patch and git-send-email to
patches@git.madduck.net.
If you'd read over the Git project's submission guidelines and adhered to them,
I'd be especially grateful.
Richard Si [Fri, 16 Jul 2021 00:21:53 +0000 (20:21 -0400)]
Don't include profiling/ to cut down sdist by ~2x (#2362)
They seem to be used as test cases for a specific region of formatting
that was slow. Now performance testing is probably something end users
won't be needing to do, so this is an easy way of reducing the sdist
size sigificantly.
Richard Si [Mon, 12 Jul 2021 20:01:38 +0000 (16:01 -0400)]
Switch `toml` TOML library for `tomli` (#2301)
toml unfortunately has a lack of maintainership issue right now. It's
evident by the fact toml only supports TOML v0.5.0. TOML v1.0.0 has
been recently released and right now Black crashes hard on its usage.
tomli is a brand new parse only TOML library. It supports TOML
v1.0.0. Although TBH we're switching to this one mostly because
pip is doing the same.
*The upper bound was included at the library maintainer's request.
Vim plugin fix string normalization option (#1869)
This commit fixes parsing of the skip-string-normalization option in vim
plugin. Originally, the plugin read the string-normalization option,
which does not exist in help (--help) and it's not respected by black
on command line.
Commit history before merge:
* fix string normalization option in vim plugin
* fix string normalization option in vim plugin
* Finish and fix patch (thanks Matt Wozniski!)
FYI: this is totally the work and the comments below of Matt (AKA godlygeek)
This fixes two entirely different problems related to how pyproject.toml
files are handled by the vim plugin.
=== Problem #1 ===
The plugin fails to properly read boolean values from pyproject.toml.
For instance, if you create this pyproject.toml:
```
[tool.black]
quiet = true
```
the Black CLI is happy with it and runs without any messages, but the
:Black command provided by this plugin fails with:
```
Traceback (most recent call last):
File "<string>", line 1, in <module>
File "<string>", line 102, in Black
File "<string>", line 150, in get_configs
File "<string>", line 150, in <dictcomp>
File "/usr/lib/python3.6/distutils/util.py", line 311, in strtobool
val = val.lower()
AttributeError: 'bool' object has no attribute 'lower'
```
That's because the value returned by the toml.load() is already a
bool, but the vim plugin incorrectly tries to convert it from a str to a bool.
The value returned by toml_config.get() was always being passed to
flag.cast(), which is a function that either converts a string to an
int or a string to a bool, depending on the flag. vim.eval()
returns integers and strings all as str, which is why we need the cast,
but that's the wrong thing to do for values that came from toml.load().
We should be applying the cast only to the return from vim.eval()
(since we know it always gives us a string), rather than casting the
value that toml.load() found - which is already the right type.
=== Problem #2 ===
The vim plugin fails to take the value for skip_string_normalization
from pyproject.toml. That's because it looks for a string_normalization
key instead of a skip_string_normalization key, thanks to this line
saying the name of the flag is string_normalization:
black/autoload/black.vim (line 25 in 05b54b8)
```
Flag(name="string_normalization", cast=strtobool),
```
and this dictcomp looking up each flag's name in the config dict:
black/autoload/black.vim (lines 148 to 151 in 05b54b8)
```
return {
flag.var_name: flag.cast(toml_config.get(flag.name, vim.eval(flag.vim_rc_name)))
for flag in FLAGS
}
```
For the second issue, I think I'd do a slightly different patch. I'd
keep the change to invert this flag's meaning and change its name that
this PR proposes, but I'd also change the handling of the
g:black_skip_string_normalization and g:black_string_normalization
variables to make it clear that g:black_skip_string_normalization is
the expected name, and g:black_string_normalization is only checked
when the expected name is unset, for backwards compatibility.
My proposed behavior is to check if g:black_skip_string_normalization
is defined and to define it if not, using the inverse of
g:black_string_normalization if that is set, and otherwise to the
default of 0. The Python code in autoload/black.vim runs later, and
will use the value of g:black_skip_string_normalization (and ignore
g:black_string_normalization; it will only be used to set
g:black_skip_string_normalization if it wasn't already set).
---
Co-authored-by: Matt Wozniski <mwozniski@bloomberg.net>
* Fix plugin/black.vim (need to up my vim game)
Co-authored-by: Matt Wozniski <godlygeek@gmail.com> Co-authored-by: Richard Si <63936253+ichard26@users.noreply.github.com> Co-authored-by: Matt Wozniski <mwozniski@bloomberg.net> Co-authored-by: Matt Wozniski <godlygeek@gmail.com>
Austin Glaser [Sat, 12 Jun 2021 19:52:49 +0000 (12:52 -0700)]
Find pyproject from vim relative to current file (#1871)
Commit history before merge:
* Find pyproject from vim relative to current file
* Merge remote-tracking branch 'upstream/main' into find-pyproject-vim
* Finish and fix this patch (thanks Matt Wozniski!)
Both the existing code and the proposed code are broken.
The vim.eval() call (whether it's vim.eval("@%") or
vim.eval("fnamemodify(getcwd(), ':t')) returns a string, and it passes
that string to find_pyproject_toml, which expects a sequence of strings,
not a single string, and - since a string is a sequence of single
character strings - it gets turned into a list of ridiculous paths. I
tested with a file called foo.py, and added a print(path_srcs) into
find_project_root, which printed out:
This does work for an unnamed buffer, too - we wind up calling
black.find_pyproject_toml(("",)), and that winds up prepending the
working directory to any relative paths, so "" just gets turned into
the current working directory.
Note that find_pyproject_toml needs to be passed a 1-tuple, not a
list, because it requires something hashable (thanks to
functools.lru_cache being used)
Co-authored-by: Matt Wozniski <mwozniski@bloomberg.net>
* I forgot the CHANGELOG entry ... again
* I'm really bad at dealing with merge conflicts sometimes
* Be more correct describing search behaviour
Co-authored-by: Austin Glaser <austin.glaser@spacex.com> Co-authored-by: Richard Si <63936253+ichard26@users.noreply.github.com> Co-authored-by: Matt Wozniski <mwozniski@bloomberg.net>
Cooper Lees [Fri, 11 Jun 2021 04:06:50 +0000 (21:06 -0700)]
Add STDIN test to primer (#2315)
* Add STDIN test to primer
- Check that out STDIN black support stays working
- Add asyncio.subprocess STDIN pip via communicate
- We just check we format python code from primer's `lib.py`
Richard Si [Wed, 9 Jun 2021 00:46:09 +0000 (20:46 -0400)]
Regression fix: leave R prefixes capitalization alone (#2285)
`black.strings.get_string_prefix` used to lowercase the extracted
prefix before returning it. This is wrong because 1) it ignores the
fact we should leave R prefixes alone because of MagicPython, and 2)
there is dedicated prefix casing handling code that fixes issue 1.
`.lower` is too naive.
This was originally fixed in 20.8b0, but was reintroduced since 21.4b0.
I also added proper prefix normalization for docstrings by using the
`black.strings.normalize_string_prefix` helper.
Some more test strings were added to make sure strings with capitalized
prefixes aren't treated differently (actually happened with my original
patch, Jelle had to point it out to me).
Felix Hildén [Tue, 8 Jun 2021 21:57:23 +0000 (00:57 +0300)]
Mention comment non-processing in documentation (#2306)
This commit adds a short section discussing the non-processing of docstrings
besides spacing improvements, mentions comment moving and links to the
AST equivalence discussion. I also added a simple spacing test for good
measure.
Commit history before merge:
* Mention comment non-processing in documentation, add spacing test
* Mention special cases for comment spacing
* Add all special cases, improve wording
Sergey Vartanov [Tue, 8 Jun 2021 21:37:34 +0000 (00:37 +0300)]
Possible fix for issue with indentation and fmt: skip (#2281)
Not sure the fix is right. Here is what I found: issue is connected
with line
first.prefix = prefix[comment.consumed :]
in `comments.py`. `first.prefix` is a prefix of the line, that ends
with `# fmt: skip`, but `comment.consumed` is the length of the
`" # fmt: skip"` string. If prefix length is greater than 14,
`first.prefix` will grow every time we apply formatting.
Cooper Lees [Mon, 7 Jun 2021 15:05:08 +0000 (08:05 -0700)]
[primer] Enable everything (#2288)
See if we pass all our repos with experimental string processing enabled.
Django probably needed:
- Ignores >= 3.8 only
We could support PEP440 version specifiers, but that would introduce the packaging module as a dependency that I'd like to avoid ... Or I could implement a poor persons version or vendor
Commit history before merge:
* [primer] Enable everything
* Add exclude extend to django CLI args for primer
* Change default timeout to from 5 to 10 mins for a primer project
* Skip string normalization for Django
* Limit Django to >= 3.8 due to := operator
Richard Si [Fri, 4 Jun 2021 01:26:21 +0000 (21:26 -0400)]
Go back to single core for test suite on CI (#2305)
The random asyncio bug is just too frequent and annoying to be
worth the speed improvements. Our test suite is already quite fast.
Random test failures hurt for 3 reasons, 1) they are discouraging for
new contributors who won't understand it's out of their control, 2)
it's annoying and time consuming to rerun the workflow, and 3) it
makes single job failures feel less important (even they should be
treated as important!).
Felix Hildén [Thu, 3 Jun 2021 20:09:41 +0000 (23:09 +0300)]
Add option to require a specific version to be running (#2300)
Closes #1246: This PR adds a new option (and automatically a toml entry, hooray for existing configuration management 🎉) to require a specific version of Black to be running.
For example: `black --required-version 20.8b -c "format = 'this'"`
Execution fails straight away if it doesn't match `__version__`.
Stefan Foulis [Tue, 1 Jun 2021 01:45:50 +0000 (03:45 +0200)]
Add `version` to github action (and rewrite the whole thing while at it) (#1940)
Commit history before merge:
* Add black_version to github action
* Merge upstream/main into this branch
* Add version support for the Black action pt.2
Since we're moving to a composite based action, quite a few changes
were made. 1) Support was added for all OSes (Windows was painful).
2) Isolation from the rest of the workflow had to be done manually
with a virtual environment.
Other noteworthy changes:
- Rewrote basically all of the logic and put it in a Python script
for easy testing (not doing it here tho cause I'm lazy and I can't
think of a reasonable way of testing it).
- Renamed `black_version` to `version` to better fit the existing
input naming scheme.
- Added support for log groups, this makes our action's output a
bit more fancy (I may or may have not added some debug output too).
* Add more to and sorta rewrite the Action's docs
Reflect compatability and gotchas.
* Add CHANGELOG entry
* Merge main into this branch
* Remove debug; address typos; clean up action.yml
Co-authored-by: Richard Si <63936253+ichard26@users.noreply.github.com>
Bryan Bugyi [Tue, 1 Jun 2021 00:57:23 +0000 (20:57 -0400)]
Correct max string length calculation when there are string operators (#2292)
PR #2286 did not fix the edge-cases (e.g. when the string is just long
enough to cause a line to be 89 characters long). This PR corrects that
mistake.
Jelle Zijlstra [Sat, 29 May 2021 16:03:08 +0000 (09:03 -0700)]
Fix path_empty() (#2275)
Behavior other than output shouldn't depend on the verbose/quiet option. As far as I can tell this currently has no visible effect, since code after this function is called handles an empty list gracefully.
Matteo Bertucci [Sat, 29 May 2021 13:56:46 +0000 (15:56 +0200)]
Issue templates: use HTML comments (#2269)
This commit makes use of HTML comments inside GitHub issue templates
to make sure that even if they aren't removed by the issue author they won't be shown
in the rendered output.
The goal is to simply make the issues less noisy by removing template messages.
Richard Si [Thu, 27 May 2021 02:04:10 +0000 (22:04 -0400)]
Fix and test docs on Windows (#2262)
There's some weird interaction between Click and
sphinxcontrib-programoutput on Windows that leads to an encoding error
during the printing of black-primer's help text.
Also symlinks aren't well supported on Windows so let's just use
includes which actually work because we now use MyST :D
Cooper Lees [Wed, 26 May 2021 12:52:09 +0000 (05:52 -0700)]
Add optional uvloop import (#2258)
* Add optional uvloop import
- If we find `uvloop` in the env for black, blackd or black-primer lets try and use it
- Add a uvloop extra install
Fixes #2257
Test:
- Add ci job to install black[uvloop] and run a primer run with uvloop
- Only with latest python (3.9)
- Will be handy to compare runtimes as a very unoffical benchmark
Mark Bell [Tue, 25 May 2021 22:43:28 +0000 (23:43 +0100)]
Removed adding a space into empty docstrings. (#2249)
Resolves #2168 by disabling the insertion of a " " when the docstring is entirely empty.
Note that this PR is focussed only on the case of empty docstrings. In particular this does not make any changes to the behaviour that a " " is inserted if a non-empty docstring begins with the quoting character. That is, black still prefers:
Felix Hildén [Tue, 25 May 2021 20:07:05 +0000 (23:07 +0300)]
Create FAQ documentation (GH-2247)
This commit creates a Frequently Asked Questions document for our users
to read. Hopefully they actually read it too. Items included are:
Black's non-API, AST safety, style stability, file discovery, Flake8
disagreements and Python 2 support. Hopefully I've got the answers
down in general.
Commit history before merge:
* Create FAQ
* Address feedback
* Move to single markdown file
* Minor wording improvements
* Add changelog entry
temeddix [Mon, 24 May 2021 02:19:03 +0000 (11:19 +0900)]
Solved Problem with Non-ASCII .gitignore Files (#2229)
* Solved Problem with non-alphabetical .gitignore files
When .gitignore file in the user's project directory contained non-alphabetical
characters(Japanese, Korean, Chinese, etc), Nothing works and printed this
weird message in the console('cp949' is the encoding for Korean characters
in this case). It even blocks VSCode's formatting from working. This commit
solves the problem.
Traceback (most recent call last):
File "c:\users\username\anaconda3\envs\project-name\lib\runpy.py", line 193, in _run_module_as_main
"__main__", mod_spec)
File "c:\users\username\anaconda3\envs\project-name\lib\runpy.py", line 85, in _run_code
exec(code, run_globals)
File "C:\Users\username\anaconda3\envs\project-name\Scripts\black.exe\__main__.py", line 7, in <module>
File "c:\users\username\anaconda3\envs\project-name\lib\site-packages\black\__init__.py", line 1056, in patched_main
main()
File "c:\users\username\anaconda3\envs\project-name\lib\site-packages\click\core.py", line 829, in __call__
return self.main(*args, **kwargs)
File "c:\users\username\anaconda3\envs\project-name\lib\site-packages\click\core.py", line 782, in main
rv = self.invoke(ctx)
File "c:\users\username\anaconda3\envs\project-name\lib\site-packages\click\core.py", line 1066, in invoke
return ctx.invoke(self.callback, **ctx.params)
File "c:\users\username\anaconda3\envs\project-name\lib\site-packages\click\core.py", line 610, in invoke
return callback(*args, **kwargs)
File "c:\users\username\anaconda3\envs\project-name\lib\site-packages\click\decorators.py", line 21, in new_func
return f(get_current_context(), *args, **kwargs)
File "c:\users\username\anaconda3\envs\project-name\lib\site-packages\black\__init__.py", line 394, in main
stdin_filename=stdin_filename,
File "c:\users\username\anaconda3\envs\project-name\lib\site-packages\black\__init__.py", line 445, in get_sources
gitignore = get_gitignore(root)
File "c:\users\username\anaconda3\envs\project-name\lib\site-packages\black\files.py", line 122, in get_gitignore
lines = gf.readlines()
UnicodeDecodeError: 'cp949' codec can't decode byte 0xb0 in position 13: illegal multibyte sequence
* Made .gitignore File Reader Detect Its Encoding
* Revert "Made .gitignore File Reader Detect Its Encoding"
Felix Hildén [Wed, 19 May 2021 19:11:37 +0000 (22:11 +0300)]
Link isort profile to Black code style isort mention (#2246)
The isort configuration currently in the Black code style document is
duplicated in Using Black with other tools document. I think it would
be better to consolidate information and simply link to the tool guide,
mentioning the easy profile in the original document.
I changed the link from isort PyPI page to Black's docs on isort
because for users it could be better to see the Black docs on why that
configuration is necessary and what isort is from Black's perspective.
Felix Hildén [Mon, 17 May 2021 18:38:43 +0000 (21:38 +0300)]
Make Prettier preserve line ending type (#2244)
Why? The default in Prettier 2.0 was
[changed](https://prettier.io/docs/en/options.html#end-of-line) from
`auto` to `LF`. This makes development on Windows awkward, because
every file is marked with changes both by Prettier and then by Git
regardless of repository line ending settings, making committing harder
than it should be.
---
Aside from that: I noticed that runnin pre-commit manually seems to add
line endings to symlink files, but they disappear when actually committing.
Don't know if that's a known.. quirk..(?) or not.
---
Commit history before merge:
* Make Prettier preserve line ending type
* Move options to .prettierrc
Hadi Alqattan [Sun, 16 May 2021 17:51:27 +0000 (20:51 +0300)]
Fix: black only respects the root gitignore. (#2225)
Commit history before merge:
Black now respects .gitignore files in all levels, not only root/.gitignore file
(apply .gitignore rules like git does).
* Fix: typo
* Fix: respect .gitignore files in all levels.
* Add: CHANGELOG note.
* Fix: TypeError: unsupported operand type(s) for +: 'NoneType' and 'PathSpec'
* Update docs.
* Fix: no parent .gitignore
* Add a comment since the if expression is a bit hard to understand
* Update tests - conver no parent .gitignore case.
* Use main's Pipfile.lock instead
The original changes in Pipfile.lock are whitespace only. The changes
turned the JSON's file indentation from 4 to 2. Effectively this
happened: `json.dumps(json.loads(old_pipfile_lock), indent=2) + "\n"`.
Just using main's Pipfile.lock instead of undoing the changes because
1) I don't know how to do that easily and quickly, and 2) there's a
merge conflict.
Co-authored-by: Richard Si <63936253+ichard26@users.noreply.github.com>
* Merge remote-tracking branch 'upstream/main' into i1730 …
Richard Si [Thu, 13 May 2021 19:30:34 +0000 (15:30 -0400)]
Add lower bound for aiohttp-cors + fix primer (#2231)
It appears sqlalchemy has recently reformatted their project with
Black 21.5b1.
Most of our dependencies have a lower bound and creating a test
environment with the oldest acceptable dependencies runs the full
Black test suite just fine. The only exception to this is aiohttp-cors.
It's unbounded and the oldest version 0.1.0 until 0.4.0 breaks the
test suite in such an old environment.
Failure with 0.1.0:
```
tests/test_blackd.py:10: in <module>
import blackd
testenv/lib/python3.8/site-packages/blackd/__init__.py:12: in <module>
import aiohttp_cors
testenv/lib/python3.8/site-packages/aiohttp_cors/__init__.py:29: in <module>
from .urldispatcher_router_adapter import UrlDistatcherRouterAdapter
testenv/lib/python3.8/site-packages/aiohttp_cors/urldispatcher_router_adapter.py:27: in <module>
class UrlDistatcherRouterAdapter(RouterAdapter):
testenv/lib/python3.8/site-packages/aiohttp_cors/urldispatcher_router_adapter.py:32: in UrlDistatcherRouterAdapter
def route_methods(self, route: web.Route):
E AttributeError: module 'aiohttp.web' has no attribute 'Route'
```
For 0.2.0:
```
tests/test_blackd.py:10: in <module>
import blackd
testenv/lib/python3.8/site-packages/blackd/__init__.py:12: in <module>
import aiohttp_cors
testenv/lib/python3.8/site-packages/aiohttp_cors/__init__.py:27: in <module>
from .cors_config import CorsConfig
testenv/lib/python3.8/site-packages/aiohttp_cors/cors_config.py:24: in <module>
from .urldispatcher_router_adapter import UrlDistatcherRouterAdapter
testenv/lib/python3.8/site-packages/aiohttp_cors/urldispatcher_router_adapter.py:27: in <module>
class UrlDistatcherRouterAdapter(AbstractRouterAdapter):
testenv/lib/python3.8/site-packages/aiohttp_cors/urldispatcher_router_adapter.py:32: in UrlDistatcherRouterAdapter
def route_methods(self, route: web.Route):
E AttributeError: module 'aiohttp.web' has no attribute 'Route'
```
For 0.3.0:
```
ERROR: Cannot install aiohttp-cors==0.3.0 and aiohttp==3.6.0 because these package versions have conflicting dependencies.
The conflict is caused by:
The user requested aiohttp==3.6.0
aiohttp-cors 0.3.0 depends on aiohttp<=0.20.2 and >=0.18.0
To fix this you could try to:
1. loosen the range of package versions you've specified
2. remove package versions to allow pip attempt to solve the dependency conflict
ERROR: ResolutionImpossible: for help visit https://pip.pypa.io/en/latest/user_guide/#fixing-conflicting-dependencies
```
Richard Si [Thu, 13 May 2021 01:28:41 +0000 (21:28 -0400)]
Modify when Test, Primer, and Documentation Build run (#2226)
- Test and Primer don't run for documentation only changes since it's
unnecessary, eating unnecessary cycles and slowing down CI since these
workflows eat up the 20 max workers limit quite easily!
- Documentation Build runs all of the time now since quite a bit of the
content depends on Black's code so even a simple 1-file change in
src/black/__init__.py may break the docs build. It's not like this is
a costly workflow anyway.
Fuzz is still running on all changes because with fuzzing, the more the
better in general. 6 or 7 jobs on a documentation only commit is much
better than 27/28 jobs anyway :p
I also found an error in our bug report issue template :)
Richard Si [Tue, 11 May 2021 18:09:33 +0000 (14:09 -0400)]
Remove useless flake8 config + test support code (#2221)
We've depended on Click 7.x ever since we broke CI systems across the
world (oops lol) and flake8-mypy was purged a fair bit back: #1867
Also remove the primer tests import in tests/test_black.py because it's
annoying when just trying to actually target tests/test_black.py tests.
`pytest -k test_black.py` doesn't do what you expect due to that import.
Richard Si [Mon, 10 May 2021 02:35:56 +0000 (22:35 -0400)]
Autogenerate black(d|-primer)? help in usage docs (#2212)
So these won't go out of date. This does mean the environment has be
setup a bit more carefully so the right version of the tool is used,
but thankfully the build environment is rebuilt on change on RTD anyway.
Also since the HTML docs are known to build fine, let's provide
downloadable HTMLzips of our docs.
This change needs RTD and GH to install Black with the [d] extra so
blackd's help can generated. While editing RTD's config file, let's
migrate the file to a non-deprecated filename.
Also I missed adding AUTHORS.md to the files key in the doc GHA config.
* Replace references to master branch
* Update .flake8 to reference docs on RTD
We're moving away from GitHub as a documentation host to only RTD because
it's makes our lives easier creating good docs. I know this link is dead right now,
but it won't be once we release a new version with the documentation reorganization
changes (which should be soon!).
Co-authored-by: Richard Si <63936253+ichard26@users.noreply.github.com>
Richard Si [Sat, 8 May 2021 19:17:38 +0000 (15:17 -0400)]
Reorganize docs v2 (GH-2174)
I know I know, this is the second reorganization of the docs. I'm not
saying the first one was bad or anything... but.. actually wait nah,
*it was bad*.
Anyway, welcome to probably my biggest commit. The main thing with this
reorganization was to introduce nesting to the documentation! Having
all of the docs be part of the main TOC was becoming too much. There
wasn't much room to expand either. Finally, the old setup required
a documentation generation step which was just annoying.
The goals of this reorganization was to:
1. Significantly restructure the docs to be discoverable and
understandable
2. Add room for further docs (like guides or contributing docs)
3. Get rid of the doc generation step (it was slow and frustrating)
4. Unblock other improvements and also just make contributing to the
docs easier
Another important change with this is that we are no longer using GitHub
as a documentation host. While GitHub does support Markdown based docs
actually pretty well, the lack of any features outside of GitHub Flavoured
Markdown is quite limiting. ReadTheDocs is just much better suited for
documentation. You can use reST, MyST, CommonMark, and all of their
great features like toctrees and admonitions.
Related to this change, we're adopting MyST as our flavour of Markdown.
MyST introduces neat syntax extensions to Markdown that pretty much
gives us the best of both worlds. The ease of use and simplicity of MD
and the flexibility and expressiveness of reST. Also recommonmark is
deprecated now. This switch was possible now we don't use GH as a docs
host. MyST docs have to be built to really be usable / pretty, so the MD
docs are going to look pretty bad on GH, but that's fine now!
Another thing that should be noted is that the README has been stripped
of most content since it was confusing. Users would read the README and
then think some feature or bug was fixed already and is available in a
release when in reality, they weren't. They were reading effectively
the latest docs without knowing.
See also: https://github.com/psf/black/issues/1759
FYI: CommonMark is a rationalized version of Markdown syntax
--
Commit history before merge:
* Switch to MyST-Parser + doc config cleanup
recommonmark is being deprecated in favour of MyST-Parser. This change
is welcomed, especially since MyST-Parser has syntax extensions for the
Commonmark standard. Effectively we get to use a language that's powerful
and expressive like ReST, but get the simplicity of Markdown.
The rest of this effort will be using some MyST features.
This reorganization efforts aims to remove as much duplication as possible.
The regeneration step once needed is gone, significantly simplifing our
Sphinx documentation configuration.
* Tell pipenv we replaced recommonmark for MyST-Parser
Also update `docs/requirements.txt`
* Delete all auto generated content
* Switch prettier for mdformat (plus a few plugins)
**FYI: THIS WAS EFFECTIVELY REVERTED, SEE THIRD TO LAST COMMIT**
prettier doesn't support MyST's syntax extensions which are going to be
used in this reorganization effort so we have to switch formatter.
Unfortanately mdformat's style is different from prettier's so time to
reformat the whole repo too.
We're excluding .github/ISSUE_TEMPLATE because I have no idea whether
its changes are safe, so let's play it safe.
* Move reference docs into a docs/contributing dir
They're for contributors of Black anyway. Also added a note in the
summary document warning about the lack of attention the reference has
been dealing with.
* Rewrite and setup the new landing page + main TOC
- add some more detail about Black's beta status
- add licensing info
- add external links in the main TOC for GitHub, PyPI, and IRC
- prepare main TOC for new structure
* Break out AUTHORS into its own file
Not only was the AUTHORS list quite long, this makes it easy to include
it in the Sphinx docs with just a simple symlink.
* Add license to docs via a simple include
Yes the document is orphaned but it is linked to in the landing page
(docs/index.rst).
* Add "The Black Code Style" section
This mostly was a restructuring commit, there has been a few updates but
not many. The main goal was to split "current style" and "planned
changes to the style that haven't happened yet" to avoid confusion.
* Add "Getting Started" page
This is basically a quick start + even more. This commit is certainly
one of most creatively involved in this effort.
* Add "Usage and Configuration" section
This commit was as much restructuring as new content. Instead of being
in one giant file, usage and configuration documentation can expand
without bloating a single file.
* Add "Integrations" section
Just a restructuring commit ...
* Add "Guides" section
This is a promising area of documentation that could easily grow in the
future, let's prepare for that!
* Add "Contributing" section
This is also another area that I expect to see significant growth in.
Contributors to Black could definitely do with some more specific docs
that clears up certain parts of our slightly confusing project (it's
only confusing because we're getting big and old!).
* Rewrite CONTRIBUTING.md to just point to RTD
* Rewrite README.md to delegate most info to RTD
* Address feedback + a lot of corrections and edits
I know I said I wanted to do these after landing this but given there's
going to be no time between this being merged and a release getting
pushed, I want these changes to make it in.
- drop the number flag for mdformat - to reduce diffs, see also:
https://mdformat.readthedocs.io/en/stable/users/style.html#ordered-lists
- the GH issue templates should be safe by mdformat, so get rid of the
exclude
- clarify our configuration position - i.e. stop claiming we don't have
many options, instead say we want as little formatting knobs as
possible
- lots and lots of punctuation, spelling, and grammar corrections (thanks
Jelle!)
- use RTD as the source for the CHANGELOG too
- visual style cleanups
- add docs about our .gitignore behaviour
- expand GHA Action docs
- clarify we want the PR number in the CHANGELOG entry
- claify Black's behaviour for with statements post Python 3.9
- italicize a bunch of "Black"s
Thank you goes to Jelle, Taneli (hukkinj1 on GH), Felix
(felix-hilden on GH), and Wouter (wbolster on GH) for the feedback!
* Merge remote-tracking branch 'upstream/master' into reorganize-docs-v2
merge conflicts suck, although these ones weren't too bad.
I consider this important enough to be worthy of a changelog entry :)
* Merge branch 'master' into reorganize-docs-v2
Co-authored-by: Łukasz Langa <lukasz@langa.pl>
* Actually let's continue using prettier
Prettier works fine for all of the default MyST syntax so let's not
rock the boat as much. Dropping the mdformat commit was merge-conflict
filled so here's additional commit instead.
* Address Cooper's, Taneli's, and Jelle's feedback
Lots of wording improvements by Cooper. Taneli suggested to disable the
enabled by default MyST syntax not supported by Prettier and I agreed.
And Jelle found one more spelling error!
Richard Si [Sat, 8 May 2021 09:34:25 +0000 (05:34 -0400)]
Speed up tests even more (#2205)
There's three optimizations in this commit:
1. Don't check if Black's output is stable or equivalant if no changes
were made in the first place. It's not like passing the same code
(for both source and actual) through black.assert_equivalent or
black.assert_stable is useful. It's not a big deal for the smaller
tests, but it eats a lot of time in tests/test_format.py since
its test cases are big. This is also closer to how Black works IRL.
2. Use a smaller file for `test_root_logger_not_used_directly` since
the logging it's checking happens during blib2to3's startup so the
file doesn't really matter.
3. If we're checking a file is formatting (i.e. test_source_is_formatted)
don't run Black over it again with `black.format_file_in_place`.
`tests/test_format.py::TestSimpleFormat.check_file` is good enough.
Łukasz Langa [Sat, 8 May 2021 09:29:47 +0000 (11:29 +0200)]
Refactor `src/black/__init__.py` into many files (#2206)
* Move string-related utility to functions to strings.py, const.py
* Move Leaf/Node-related functionality to nodes.py
* Move comment-related functions to comments.py
* Move caching to cache.py and Mode/TargetVersion/Feature to mode.py
* Move some leftover functions to nodes.py, comments.py, strings.py
* Add missing files to source list for test runs
* Move line-related functionality into lines.py, brackets into brackets.py
* Move transformers to trans.py
* Move file handling, output, parsing, concurrency, debug, and report
* Move two more functions to nodes.py
* Add CHANGES
* Add numeric.py
* Add linegen.py
* More docstrings
* Include new files in tests
Richard Si [Fri, 7 May 2021 14:41:55 +0000 (10:41 -0400)]
Speed up test suite via distributed testing (#2196)
* Speed up test suite via distributed testing
Since we now run the test suite twice, one with Python 2 and another
without, full test runs are getting pretty slow. Let's try to
fix that with parallization.
Also use verbose mode on CI since more logs is usually better since
getting more is quite literally impossible.
The main issue we'll face with this is we'll hit
https://github.com/pytest-dev/pytest-xdist/issues/620 sometimes
(although pretty rarely). I suppose we can test this and see if how bad
this bug is for us, and revert if necessary down the line.
Łukasz Langa [Fri, 7 May 2021 14:33:36 +0000 (16:33 +0200)]
Mark blackd tests with the `blackd` optional marker (#2204)
This is a follow-up of #2203 that uses a pytest marker instead of a bunch of
`skipUnless`. Similarly to the Python 2 tests, they are running by default and
will crash on an unsuspecting contributor with missing dependencies. This is
by design, we WANT contributors to test everything. Unless we actually don't
and then we can run:
Kaleb Barrett [Fri, 7 May 2021 12:54:21 +0000 (07:54 -0500)]
Do not use gitignore if explicitly passing excludes (#2170)
Closes #2164.
Changes behavior of how .gitignore is handled. With this change, the rules in .gitignore are only used as a fallback if no exclusion rule is explicitly passed on the command line or in pyproject.toml. Previously they were used regardless if explicit exclusion rules were specified, preventing any overriding of .gitignore rules.
Those that depend only on .gitignore for their exclusion rules will not be affected. Those that use both .gitignore and exclude will find that exclude will act more like actually specifying exclude and not just another extra-excludes. If the previous behavior was desired, they should move their rules from exclude to extra-excludes.
Shota Ray Imaki [Thu, 6 May 2021 02:25:43 +0000 (11:25 +0900)]
Simplify GitHub Action entrypoint (#2119)
This commit simplifies entrypoint.sh for GitHub Actions by removing
duplication of args and black_args (cf. #1909).
The reason why #1909 uses the input id black_args is to avoid an overlap
with args, but this naming seems redundant. So let me suggest option
and src, which are consistent with CLI. Backward compatibility is
guaranteed; Users can still use black_args as well.
Commit history pre-merge:
* Simplify GitHub Action entrypoint (#1909)
* Fix prettier
* Emit a warning message when `black_args` is used
This deprecation should be visible in GitHub Action's UI now.
Co-authored-by: Shota Ray Imaki <shota.imaki.0801@gmail.com> Co-authored-by: Richard Si <63936253+ichard26@users.noreply.github.com>
Cooper Lees [Wed, 5 May 2021 15:33:23 +0000 (08:33 -0700)]
Enable ` --experimental-string-processing` on most primer projects (#2184)
* Enable ` --experimental-string-processing` on all primer projects
- We want to make this default so need to test it more
- Fixed splat/star bug in extending black args for each project
Richard Si [Tue, 4 May 2021 08:47:59 +0000 (04:47 -0400)]
Drop Travis CI and migrate Coveralls (#2186)
Travis CI for Open Source is shutting down in a few weeks so the queue
for jobs is insane due to lower resources. I'm 99.99% sure we don't need
it as our Test, Lint, Docs, Upload / Package, Primer, and Fuzz workflows
are all on GitHub Actions. So even though we *can* migrate to the .com
version with its 1000 free Linux minutes(?), I don't think we need to.
more information here:
- https://blog.travis-ci.com/oss-announcement
- https://blog.travis-ci.com/2020-11-02-travis-ci-new-billing
- https://docs.travis-ci.com/user/migrate/open-source-repository-migration
This commit does the following:
- delete the Travis CI configuration
- add to the GHA test workflows so coverage continues to be recorded
- tweaked coverage configuration so this wouldn't break
- remove any references to Travis CI in the docs (i.e. readme + sphinx
docs)
Regarding the Travis CI to GitHub Actions Coveralls transition, the
official action doesn't support the coverage files produced by coverage.py
unfornately. Also no, I don't really know what I am doing so don't @ me
if this breaks :p (well you can, but don't expect me to be THAT useful).
The Coveralls setup has two downfalls AFAIK:
- Only Linux runs are used because AndreMiras/coveralls-python-action
only supports Linux. Although this isn't a big issue since the Travis
Coveralls configuration only used Linux data too.
- Pull requests from an internal branch (i.e. one on psf/black) will be
marked as a push coverage build by Coveralls since our anti-duplicate-
workflows system runs under the push even for such cases.
projects / etc / vim.git / log