Fixes#1360, where an invalid config file causes a return/exit code of 1. This
change means this case is caught earlier, treated like any other bad
parameters, and results in an exit code of 2.
Co-authored-by: Toby Fleming <tobywf@users.noreply.github.com>
`split("\n")` includes a final empty element `""` if the final line
ends with `\n` (as it should for POSIX-compliant text files), which
then became an extra `"\n"`.
`splitlines()` solves that, but there's a caveat, as it will split
on other types of line breaks too (like `\r`), which may not be
desired.
Fixes#526.
`type: ignore` shouldn't block collapsing a line, since it will still
apply fine to the merged line. This prevents an issue where a reformat
causes it to shift lines and then be merged on a subsequent pass.
There is a downside to this, which is that it can cause a `type:
ignore` to apply to more code than was originally intended. There
might be a way to apply this in a more limited situation, but I'm not
sure what it is.
Fixes#1061.
* Make most of blib2to3 directly typed and mypyc-compatible
This used a combination of retype and pytype's merge-pyi to do the
initial merges of the stubs, which then required manual tweaking to
make actually typecheck and work with mypyc.
Co-authored-by: Sanjit Kalapatapu <sanjitkal@gmail.com>
Co-authored-by: Michael J. Sullivan <sully@msully.net>
* Make black able to compile and run with mypyc
The changes made fall into a couple categories:
* Fixing actual type mistakes that slip through the cracks
* Working around a couple mypy bugs (the most annoying of which being
that we need to add type annotations in a number of places where
variables are initialized to None)
Co-authored-by: Sanjit Kalapatapu <sanjitkal@gmail.com>
Co-authored-by: Michael J. Sullivan <sully@msully.net>
The old behavior would detect the existence of a `# fmt: on` in a leaf
node's comment prefix and immediately mark the node as formatting-on,
even if a subsequent `# fmt: off` in the same comment prefix would turn
it back off. This change modifies that logic to track the state through
the entire prefix and take the final state.
Note that this does not fully solve on/off behavior, since any _comment_
lines between the off/on are still formatted. We may need to add
virtual leaf nodes to truly solve that. I will leave that for a separate
commit/PR.
Fixes#1005
Fixes#1042 (and probably #1044 which looks like the same thing).
The issue with the "obviously unnecessary" parentheses that #850 removed is that sometimes they're necessary to help Black fit something in one line. I didn't see an obvious solution that still removes the parens #850 was intended to remove, so let's back out this change for now in the interest of unblocking a release.
This PR also adds a test adapted from the failing example in #1042, so that if we try to reapply the #850 change we don't break the same case again.
* Blacken .py files in blib2to3
This is in preparation for adding type annotations to blib2to3 in
order to compiling it with mypyc (#1009, which I can rebase on top of
this).
To enforce that it stays blackened, I just cargo-culted the existing
test code used for validating formatting. It feels pretty clunky now,
though, so I can abstract the common logic out into a helper if that
seems better. (But error messages might be less clear then?)
* Tidy up the tests
It turns out we also need to handle invisible *left* parens added at
the *start* of a line. Refactor `contains_unsplittable_type_ignore` to
handle this more cleanly.
* add test for special unicode symbol which usual re can not process correctly
add regex lib which supports unicode 12.1.0 standard
replace re usage in project in favor to regex
* #455 fix dependency
In #1040 I had convinced myself that the type ignore logic didn't
need anything like the ignored_ids from the type comment logic, but I
was wrong, and we do.
We hit these cases in practice a bunch.
The code introduced in #1027 to detect whether a type comment appeared
after a regular comment in a Line would spuriously misfire when a leaf
was in the comments dict but had an empty list of comments. This can
occur as an artifact of how comments on trailing commas are handled,
it seems.
(This was discovered trying to test black out on mypy.)
Type comments only apply if they are the first comment on the line,
which means that allowing them to be pushed behind a regular comment
when joining lines is a semantic change (and, indeed, one that black
catches and fails on).
* Parse `:=` properly
* never unwrap parenthesis around `:=`
* When checking for AST-equivalence, use `ast` instead of `typed-ast` when running on python >=3.8
* Assume code that uses `:=` is at least 3.8
Modified maybe_remove_trailing_comma to remove trailing commas for
typedarglists (in addition to arglists), and updated line split logic
to ensure that all lines in a function definition that contain only one
arg have a trailing comma.
Based on the feedback in
https://github.com/python/black/pull/845#issuecomment-490622711
- Remove TokenizerConfig, and add a field to Grammar instead.
- Pass the Grammar to the tokenizer.
- Rename `ASYNC_IS_RESERVED_KEYWORD` to `ASYNC_KEYWORDS` and
`ASYNC_IS_VALID_IDENTIFIER` to `ASYNC_IDENTIFIERS`.
Fixes#593
I looked into this bug with @ambv and @carljm, and we reached the
conclusion was that it's not possible for the tokenizer to determine if
async/await is a keyword inside all possible generators without breaking
the grammar for older versions of Python.
Instead, we introduce a new tokenizer mode for Python 3.7+ that will
cause all async/await instances to get parsed as a reserved keyword,
which should fix async/await inside generators.
... to stop it from spamming the log when black is used as a library in another
python application.
When used indirectly by black the logger initiated in `driver.py` will emit
thousands of debug messages making the debug level of the root logger virtually
useless. By getting a named logger instead the verbosity of logging from this
module can easily be controlled by setting its log level.
Fixes#715
Fixes#305
- Run separate jobs for mypy, self-formatting, flake8, and test runs.
- Don't run flake8 in 3.8 because it is broken (and we can't really expect flake8 to always keep up with 3.8 development).
- Fix unused variable in test
This is a new syntax added in python3.7, so black can't verify that reformatting will not change the ast unless black itself is run with 3.7. We'll need to change the error message black gives in this case. @ambv any ideas?
Fixes#125.
Fixes#452
I ended up making a couple of other normalizations to numeric literals
too (lowercase everything, don't allow leading or trailing . in floats,
remove redundant + sign in exponent). I don't care too much about those,
so I'm happy to change the behavior there.
For reference, here is Python's grammar for numeric literals:
https://docs.python.org/3/reference/lexical_analysis.html#numeric-literals
These 2 options allow you to pass in regular expressions that determine
whether files/directories are included or excluded in the recursive file
search.
Fixes#270
I documented the new behavior, added it to the change log, greatly expanded
tests, added support for inner defs and classes, and added Luka to ACKS.
Fixes#196
If an expression starts or ends with a bracket and only contains a single
delimiter, don't wrap it in additional optional parentheses. We can use the
brackets for the split.
Fixes#177Fixes#193
Fixes#140
Note: this is an evolution but the end result needs to be different. See
cantfit.py for some good examples on bad formatting caused by this change.
Fixes a pathological situation where if a function signature used a trailing
comma but was later reformatted to a single line (with the trailing comma
removed), Black would change its mind whether a file is Python
3.6-compatible between runs.
Just executing ``black`` without any argument does not print any message
to stdout or stderr. It's rather confusing, because the user doesn't
know what happened.
In ``len(sources) == 0`` case, black now prints ``No paths given. Nothing to
do``.
Signed-off-by: Christian Heimes <christian@python.org>
Black will cache already formatted files using their file size and
modification timestamp. The cache is per-user and will always be used
unless Black is used with --diff or with code provided via standard
input.
This support isn't *exactly* right per PEP 3131 as the regex engine is a bit
too limited for that and I didn't want to spend time on Other_ID_Start and
Other_ID_Continue unless they're actually needed.
Hopefully this doesn't slow it down too much.
In raw strings, a single backslash means a literal backslash. It is also used to escape quotes if it precedes them. This means it is impossible to change the quote type for strings that contain an unescaped version of the other quote type.
Fixes#100
Black cannot currently support this form due to its generator-based nature.
This is mostly a problem for existing `# yapf: disable` usage as trailing
comment.
Fixes#95
The default behaviour is that now all lines break *before* delimiters,
instead of afterwards. The special cases for this are commas and
behaviour around args.
Resolves#73
* Normalize string quotes
Convert single-quoted strings to double-quoted. Convert triple single-quoted strings to triple double-quoted. Do not touch any strings where conversion would increase the number of backslashes.
Fixes#51.
* reformat Black itself
Now Black properly splits standalone comments within bracketed expressions.
They are treated as another type of split instead of being bolted on with
whitespace prefixes.
A related fix: now multiple comments might appear after a given leaf.
Fixes#22
Being able to format code by piping it through the formatter makes it much easier to integrate with tools like google/vim-codefmt or Chiel92/vim-autoformat.
This makes it consistent with removing the trailing comma when multiple
arguments to a call fit in a single line. It also makes it a tiny bit more
likely that an expression will fit a line that didn't use to.
Trailing commas after * or ** in a function signature are only safe for Python 3.6
code. So now Black checks whether the file was already Python 3.6 to begin
with. If so, trailing commas are used in such cases. Otherwise, they're not.
When * and ** don't appear in a function signature, the trailing comma is
always safe.
Fixes#8
