This implements formatting ExprTuple, including magic trailing comma. I
intentionally didn't change the settings mechanism but just added a
dummy global const flag.
Besides the snapshots, I added custom breaking/joining tests and a
deeply nested test case. The diffs look better than previously, proper
black compatibility depends on parentheses handling.
---------
Co-authored-by: Micha Reiser <micha@reiser.io>
* A basic StmtAssign formatter and better dummies for expressions
The goal of this PR was formatting StmtAssign since many nodes in the black tests (and in python in general) are after an assignment. This caused unstable formatting: The spacing of power op spacing depends on the type of the two involved expressions, but each expression was formatted as dummy string and re-parsed as a ExprName, so in the second round the different rules of ExprName were applied, causing unstable formatting.
This PR does not necessarily bring us closer to black's style, but it unlocks a good porting of black's test suite and is a basis for implementing the Expr nodes.
* fmt
* Review
<!--
Thank you for contributing to Ruff! To help us out with reviewing, please consider the following:
- Does this pull request include a summary of the change? (See below.)
- Does this pull request include a descriptive title?
- Does this pull request include references to any relevant issues?
-->
## Summary
This PR replaces the `verbatim_text` builder with a `not_yet_implemented` builder that emits `NOT_YET_IMPLEMENTED_<NodeKind>` for not yet implemented nodes.
The motivation for this change is that partially formatting compound statements can result in incorrectly indented code, which is a syntax error:
```python
def func_no_args():
a; b; c
if True: raise RuntimeError
if False: ...
for i in range(10):
print(i)
continue
```
Get's reformatted to
```python
def func_no_args():
a; b; c
if True: raise RuntimeError
if False: ...
for i in range(10):
print(i)
continue
```
because our formatter does not yet support `for` statements and just inserts the text from the source.
## Downsides
Using an identifier will not work in all situations. For example, an identifier is invalid in an `Arguments ` position. That's why I kept `verbatim_text` around and e.g. use it in the `Arguments` formatting logic where incorrect indentations are impossible (to my knowledge). Meaning, `verbatim_text` we can opt in to `verbatim_text` when we want to iterate quickly on nodes that we don't want to provide a full implementation yet and using an identifier would be invalid.
## Upsides
Running this on main discovered stability issues with the newline handling that were previously "hidden" because of the verbatim formatting. I guess that's an upside :)
## Test Plan
None?
<!--
Thank you for contributing to Ruff! To help us out with reviewing, please consider the following:
- Does this pull request include a summary of the change? (See below.)
- Does this pull request include a descriptive title?
- Does this pull request include references to any relevant issues?
-->
## Summary
This issue fixes the removal of empty lines between a leading comment and the previous statement:
```python
a = 20
# leading comment
b = 10
```
Ruff removed the empty line between `a` and `b` because:
* The leading comments formatting does not preserve leading newlines (to avoid adding new lines at the top of a body)
* The `JoinNodesBuilder` counted the lines before `b`, which is 1 -> Doesn't insert a new line
This is fixed by changing the `JoinNodesBuilder` to count the lines instead *after* the last node. This correctly gives 1, and the `# leading comment` will insert the empty lines between any other leading comment or the node.
## Test Plan
I added a new test for empty lines.
* Add Format for Stmt
* Implement basic module formatting
This implements formatting each statement in a module with a hard line break in between, so that we can start formatting statements.
Basic testing is done by the snapshots
* Use dummy verbatim formatter for all nodes
* Use new formatter infrastructure in CLI and test
* Expose the new formatter in the CLI
* Merge import blocks
This just re-formats all the `.py.expect` files with Black, both to add a trailing newline and be doubly-certain that they're correctly formatted.
I also ensured that we add a hard line break after each statement, and that we avoid including an extra newline in the generated Markdown (since the code should contain the exact expected newlines).
This PR changes the testing infrastructure to run all black tests and:
* Pass if Ruff and Black generate the same formatting
* Fail and write a markdown snapshot that shows the input code, the differences between Black and Ruff, Ruffs output, and Blacks output
This is achieved by introducing a new `fixture` macro (open to better name suggestions) that "duplicates" the attributed test for every file that matches the specified glob pattern. Creating a new test for each file over having a test that iterates over all files has the advantage that you can run a single test, and that test failures indicate which case is failing.
The `fixture` macro also makes it straightforward to e.g. setup our own spec tests that test very specific formatting by creating a new folder and use insta to assert the formatted output.
# Summary
This PR contains the code for the autoformatter proof-of-concept.
## Crate structure
The primary formatting hook is the `fmt` function in `crates/ruff_python_formatter/src/lib.rs`.
The current formatter approach is outlined in `crates/ruff_python_formatter/src/lib.rs`, and is structured as follows:
- Tokenize the code using the RustPython lexer.
- In `crates/ruff_python_formatter/src/trivia.rs`, extract a variety of trivia tokens from the token stream. These include comments, trailing commas, and empty lines.
- Generate the AST via the RustPython parser.
- In `crates/ruff_python_formatter/src/cst.rs`, convert the AST to a CST structure. As of now, the CST is nearly identical to the AST, except that every node gets a `trivia` vector. But we might want to modify it further.
- In `crates/ruff_python_formatter/src/attachment.rs`, attach each trivia token to the corresponding CST node. The logic for this is mostly in `decorate_trivia` and is ported almost directly from Prettier (given each token, find its preceding, following, and enclosing nodes, then attach the token to the appropriate node in a second pass).
- In `crates/ruff_python_formatter/src/newlines.rs`, normalize newlines to match Black’s preferences. This involves traversing the CST and inserting or removing `TriviaToken` values as we go.
- Call `format!` on the CST, which delegates to type-specific formatter implementations (e.g., `crates/ruff_python_formatter/src/format/stmt.rs` for `Stmt` nodes, and similar for `Expr` nodes; the others are trivial). Those type-specific implementations delegate to kind-specific functions (e.g., `format_func_def`).
## Testing and iteration
The formatter is being developed against the Black test suite, which was copied over in-full to `crates/ruff_python_formatter/resources/test/fixtures/black`.
The Black fixtures had to be modified to create `[insta](https://github.com/mitsuhiko/insta)`-compatible snapshots, which now exist in the repo.
My approach thus far has been to try and improve coverage by tackling fixtures one-by-one.
## What works, and what doesn’t
- *Most* nodes are supported at a basic level (though there are a few stragglers at time of writing, like `StmtKind::Try`).
- Newlines are properly preserved in most cases.
- Magic trailing commas are properly preserved in some (but not all) cases.
- Trivial leading and trailing standalone comments mostly work (although maybe not at the end of a file).
- Inline comments, and comments within expressions, often don’t work -- they work in a few cases, but it’s one-off right now. (We’re probably associating them with the “right” nodes more often than we are actually rendering them in the right place.)
- We don’t properly normalize string quotes. (At present, we just repeat any constants verbatim.)
- We’re mishandling a bunch of wrapping cases (if we treat Black as the reference implementation). Here are a few examples (demonstrating Black's stable behavior):
```py
# In some cases, if the end expression is "self-closing" (functions,
# lists, dictionaries, sets, subscript accesses, and any length-two
# boolean operations that end in these elments), Black
# will wrap like this...
if some_expression and f(
b,
c,
d,
):
pass
# ...whereas we do this:
if (
some_expression
and f(
b,
c,
d,
)
):
pass
# If function arguments can fit on a single line, then Black will
# format them like this, rather than exploding them vertically.
if f(
a, b, c, d, e, f, g, ...
):
pass
```
- We don’t properly preserve parentheses in all cases. Black preserves parentheses in some but not all cases.
