This PR adjusts the logic for skipping formatting so that a `fmt: skip`
can affect multiple statements if they lie on the same line.
Specifically, a `fmt: skip` comment will now suppress all the statements
in the suite in which it appears whose range intersects the line
containing the skip directive. For example:
```python
x=[
'1'
];x=2 # fmt: skip
```
remains unchanged after formatting.
(Note that compound statements are somewhat special and were handled in
a previous PR - see #20633).
Closes#17331 and #11430.
Simplest to review commit by commit - the key diffs of interest are the
commit introducing the core logic, and the diff between the snapshots
introduced in the last commit (compared to the second commit).
# Implementation
On `main` we format a suite of statements by iterating through them. If
we meet a statement with a leading or trailing (own-line)`fmt: off`
comment, then we suppress formatting until we meet a `fmt: on` comment.
Otherwise we format the statement using its own formatting rule.
How are `fmt: skip` comments handled then? They are handled internally
to the formatting of each statement. Specifically, calling `.fmt` on a
statement node will first check to see if there is a trailing,
end-of-line `fmt: skip` (or `fmt: off`/`yapf: off`), and if so then
write the node with suppressed formatting.
In this PR we move the responsibility for handling `fmt: skip` into the
formatting logic of the suite itself. This is done as follows:
- Before beginning to format the suite, we do a pass through the
statements and collect the data of ranges with skipped formatting. More
specifically, we create a map with key given by the _first_ skipped
statement in a block and value a pair consisting of the _last_ skipped
statement and the _range_ to write verbatim.
- We iterate as before, but if we meet a statement that is a key in the
map constructed above, we pause to write the associated range verbatim.
We then advance the iterator to the last statement in the block and
proceed as before.
## Addendum on range formatting
We also had to make some changes to range formatting in order to support
this new behavior. For example, we want to make sure that
```python
<RANGE_START>x=1<RANGE_END>;x=2 # fmt: skip
```
formats verbatim, rather than becoming
```python
x = 1;x=2 # fmt: skip
```
Recall that range formatting proceeds in two steps:
1. Find the smallest enclosing node containing the range AND that has
enough info to format the range (so it may be larger than you think,
e.g. a docstring has enclosing node given by the suite, not the string
itself.)
2. Carve out the formatted range from the result of formatting that
enclosing node.
We had to modify (1), since the suite knows how to format skipped nodes,
but nodes may not "know" they are skipped. To do this we altered the
`visit_body` bethod of the `FindEnclosingNode` visitor: now we iterate
through the statements and check for skipped ranges intersecting the
format range. If we find them, we return without descending. The result
is to consider the statement containing the suite as the enclosing node
in this case.
Resolves#21892
## Summary
This PR updates `docs/editors/features.md` to clarify that Jupyter
Notebooks are now included by default as of version 0.6.0.
## Summary
This appears to have been a copy/paste error from the list example, as
the subscript is not present in the original next/iter example only in
the case where the error case is shown. While in the specific example
code the subscript actually has no effect, it does make the example
slightly confusing.
Consider the following variations, first the example from the docs
unchanged and second the same code but not hitting the intended error
case (due to using a non-empty collection):
```console
$ python3 -c 'next(iter(range(0)))[0]'
Traceback (most recent call last):
File "<string>", line 1, in <module>
StopIteration
$ python3 -c 'next(iter(range(1)))[0]'
Traceback (most recent call last):
File "<string>", line 1, in <module>
TypeError: 'int' object is not subscriptable
```
## Test Plan
Not directly tested, however see inline snippets above.
## Summary
- Adds new RUF103 and RUF104 diagnostics for invalid and unmatched
suppression comments
- Reports RUF100 for any unused range suppression
- Reports RUF102 for range suppression comment with invalid rule codes
- Reports RUF103 for range suppression comment with invalid suppression syntax
- Reports RUF104 diagnostics for any unmatched range suppression comment (disable w/o enable)
## Test Plan
Updated snapshots from test cases with unmatched suppression comments
Issue #3711Fixes#21878Fixes#21875
This PR implements a modification (in preview) to fluent formatting for
method chains: We break _at_ the first call instead of _after_.
For example, we have the following diff between `main` and this PR (with
`line-length=8` so I don't have to stretch out the text):
```diff
x = (
- df.merge()
+ df
+ .merge()
.groupby()
.agg()
.filter()
)
```
## Explanation of current implementation
Recall that we traverse the AST to apply formatting. A method chain,
while read left-to-right, is stored in the AST "in reverse". So if we
start with something like
```python
a.b.c.d().e.f()
```
then the first syntax node we meet is essentially `.f()`. So we have to
peek ahead. And we actually _already_ do this in our current fluent
formatting logic: we peek ahead to count how many calls we have in the
chain to see whether we should be using fluent formatting or now.
In this implementation, we actually _record_ this number inside the enum
for `CallChainLayout`. That is, we make the variant `Fluent` hold an
`AttributeState`. This state can either be:
- The number of call-like attributes preceding the current attribute
- The state `FirstCallOrSubscript` which means we are at the first
call-like attribute in the chain (reading from left to right)
- The state `BeforeFirstCallOrSubscript` which means we are in the
"first group" of attributes, preceding that first call.
In our example, here's what it looks like at each attribute:
```
a.b.c.d().e.f @ Fluent(CallsOrSubscriptsPreceding(1))
a.b.c.d().e @ Fluent(CallsOrSubscriptsPreceding(1))
a.b.c.d @ Fluent(FirstCallOrSubscript)
a.b.c @ Fluent(BeforeFirstCallOrSubscript)
a.b @ Fluent(BeforeFirstCallOrSubscript)
```
Now, as we descend down from the parent expression, we pass along this
little piece of state and modify it as we go to track where we are. This
state doesn't do anything except when we are in `FirstCallOrSubscript`,
in which case we add a soft line break.
Closes#8598
---------
Co-authored-by: Brent Westbrook <36778786+ntBre@users.noreply.github.com>
Summary
--
Following #8179, we now format long lambda expressions a bit more like
Black, preferring to keep long parameter lists on a single line, but we
go one step further to break the body itself across multiple lines and
parenthesize it if it's still too long. This PR documents both the
stable deviation that breaks parameters across multiple lines, and the
new preview deviation that breaks the body instead.
I also fixed a couple of typos in the section immediately above my
addition.
Test Plan
--
I tested all of the snippets here against `main` for the preview
behavior, our playground for the stable behavior, and Black's playground
for their behavior
## Summary
The reference to the pre-commit hook inside the tutorial was to the
legacy alias `ruff` instead of the current `ruff-check`.
Ref: https://github.com/astral-sh/ruff-pre-commit/pull/124
## Test Plan
Not applicable.
Running `eglot-format` in buffers not managed by Eglot causes a
`jsonrpc-error` in Emacs 30. It may also display a
`documentFormattingProvider` warning when the server does not support
formatting. Add checks for both.
## Summary
This PR uses the new `Diagnostic` type for rendering formatter
diagnostics. This allows the formatter to inherit all of the output
formats already implemented in the linter and ty. For example, here's
the new `full` output format, with the formatting diff displayed using
the same infrastructure as the linter:
<img width="592" height="364" alt="image"
src="https://github.com/user-attachments/assets/6d09817d-3f27-4960-aa8b-41ba47fb4dc0"
/>
<details><summary>Resolved TODOs</summary>
<p>
~~There are several limitiations/todos here still, especially around the
`OutputFormat` type~~:
- [x] A few literal `todo!`s for the remaining `OutputFormat`s without
matching `DiagnosticFormat`s
- [x] The default output format is `full` instead of something more
concise like the current output
- [x] Some of the output formats (namely JSON) have information that
doesn't make much sense for these diagnostics
The first of these is definitely resolved, and I think the other two are
as well, based on discussion on the design document. In brief, we're
okay inheriting the default `OutputFormat` and can separate the global
option into `lint.output-format` and `format.output-format` in the
future, if needed; and we're okay including redundant information in the
non-human-readable output formats.
My last major concern is with the performance of the new code, as
discussed in the `Benchmarks` section below.
A smaller question is whether we should use `Diagnostic`s for formatting
errors too. I think the answer to this is yes, in line with changes
we're making in the linter too. I still need to implement that here.
</p>
</details>
<details><summary>Benchmarks</summary>
<p>
The values in the table are from a large benchmark on the CPython 3.10
code
base, which involves checking 2011 files, 1872 of which need to be
reformatted.
`stable` corresponds to the same code used on `main`, while
`preview-full` and
`preview-concise` use the new `Diagnostic` code gated behind `--preview`
for the
`full` and `concise` output formats, respectively. `stable-diff` uses
the
`--diff` to compare the two diff rendering approaches. See the full
hyperfine
command below for more details. For a sense of scale, the `stable`
output format
produces 1873 lines on stdout, compared to 855,278 for `preview-full`
and
857,798 for `stable-diff`.
| Command | Mean [ms] | Min [ms] | Max [ms] | Relative |
|:------------------|--------------:|---------:|---------:|-------------:|
| `stable` | 201.2 ± 6.8 | 192.9 | 220.6 | 1.00 |
| `preview-full` | 9113.2 ± 31.2 | 9076.1 | 9152.0 | 45.29 ± 1.54 |
| `preview-concise` | 214.2 ± 1.4 | 212.0 | 217.6 | 1.06 ± 0.04 |
| `stable-diff` | 3308.6 ± 20.2 | 3278.6 | 3341.8 | 16.44 ± 0.56 |
In summary, the `preview-concise` diagnostics are ~6% slower than the
stable
output format, increasing the average runtime from 201.2 ms to 214.2 ms.
The
`full` preview diagnostics are much more expensive, taking over 9113.2
ms to
complete, which is ~3x more expensive even than the stable diffs
produced by the
`--diff` flag.
My main takeaways here are:
1. Rendering `Edit`s is much more expensive than rendering the diffs
from `--diff`
2. Constructing `Edit`s actually isn't too bad
### Constructing `Edit`s
I also took a closer look at `Edit` construction by modifying the code
and
repeating the `preview-concise` benchmark and found that the main issue
is
constructing a `SourceFile` for use in the `Edit` rendering. Commenting
out the
`Edit` construction itself has basically no effect:
| Command | Mean [ms] | Min [ms] | Max [ms] | Relative |
|:----------|------------:|---------:|---------:|------------:|
| `stable` | 197.5 ± 1.6 | 195.0 | 200.3 | 1.00 |
| `no-edit` | 208.9 ± 2.2 | 204.8 | 212.2 | 1.06 ± 0.01 |
However, also omitting the source text from the `SourceFile`
construction
resolves the slowdown compared to `stable`. So it seems that copying the
full
source text into a `SourceFile` is the main cause of the slowdown for
non-`full`
diagnostics.
| Command | Mean [ms] | Min [ms] | Max [ms] | Relative |
|:-----------------|------------:|---------:|---------:|------------:|
| `stable` | 202.4 ± 2.9 | 197.6 | 207.9 | 1.00 |
| `no-source-text` | 202.7 ± 3.3 | 196.3 | 209.1 | 1.00 ± 0.02 |
### Rendering diffs
The main difference between `stable-diff` and `preview-full` seems to be
the diffing strategy we use from `similar`. Both versions use the same
algorithm, but in the existing
[`CodeDiff`](https://github.com/astral-sh/ruff/blob/main/crates/ruff_linter/src/source_kind.rs#L259)
rendering for the `--diff` flag, we only do line-level diffing, whereas
for `Diagnostic`s we use `TextDiff::iter_inline_changes` to highlight
word-level changes too. Skipping the word diff for `Diagnostic`s closes
most of the gap:
| Command | Mean [s] | Min [s] | Max [s] | Relative |
|:---|---:|---:|---:|---:|
| `stable-diff` | 3.323 ± 0.015 | 3.297 | 3.341 | 1.00 |
| `preview-full` | 3.654 ± 0.019 | 3.618 | 3.682 | 1.10 ± 0.01 |
(In some repeated runs, I've seen as small as a ~5% difference, down
from 10% in the table)
This doesn't actually change any of our snapshots, but it would
obviously change the rendered result in a terminal since we wouldn't
highlight the specific words that changed within a line.
Another much smaller change that we can try is removing the deadline
from the `iter_inline_changes` call. It looks like there's a fair amount
of overhead from the default 500 ms deadline for computing these, and
using `iter_inline_changes(op, None)` (`None` for the optional deadline
argument) improves the runtime quite a bit:
| Command | Mean [s] | Min [s] | Max [s] | Relative |
|:---|---:|---:|---:|---:|
| `stable-diff` | 3.322 ± 0.013 | 3.298 | 3.341 | 1.00 |
| `preview-full` | 5.296 ± 0.030 | 5.251 | 5.366 | 1.59 ± 0.01 |
<hr>
<details><summary>hyperfine command</summary>
```shell
cargo build --release --bin ruff && hyperfine --ignore-failure --warmup 10 --export-markdown /tmp/table.md \
-n stable -n preview-full -n preview-concise -n stable-diff \
"./target/release/ruff format --check ./crates/ruff_linter/resources/test/cpython/ --no-cache" \
"./target/release/ruff format --check ./crates/ruff_linter/resources/test/cpython/ --no-cache --preview --output-format=full" \
"./target/release/ruff format --check ./crates/ruff_linter/resources/test/cpython/ --no-cache --preview --output-format=concise" \
"./target/release/ruff format --check ./crates/ruff_linter/resources/test/cpython/ --no-cache --diff"
```
</details>
</p>
</details>
## Test Plan
Some new CLI tests and manual testing
