Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Keep trailing part comment associated with its part #15378

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
31 changes: 29 additions & 2 deletions crates/ruff_python_ast/src/expression.rs
Original file line number Diff line number Diff line change
Expand Up @@ -407,13 +407,13 @@ pub enum StringLike<'a> {
FString(&'a ast::ExprFString),
}

impl StringLike<'_> {
impl<'a> StringLike<'a> {
pub const fn is_fstring(self) -> bool {
matches!(self, Self::FString(_))
}

/// Returns an iterator over the [`StringLikePart`] contained in this string-like expression.
pub fn parts(&self) -> StringLikePartIter<'_> {
pub fn parts(&self) -> StringLikePartIter<'a> {
match self {
StringLike::String(expr) => StringLikePartIter::String(expr.value.iter()),
StringLike::Bytes(expr) => StringLikePartIter::Bytes(expr.value.iter()),
Expand All @@ -429,6 +429,14 @@ impl StringLike<'_> {
Self::FString(ExprFString { value, .. }) => value.is_implicit_concatenated(),
}
}

pub const fn as_expression_ref(self) -> ExpressionRef<'a> {
match self {
StringLike::String(expr) => ExpressionRef::StringLiteral(expr),
StringLike::Bytes(expr) => ExpressionRef::BytesLiteral(expr),
StringLike::FString(expr) => ExpressionRef::FString(expr),
}
}
}

impl<'a> From<&'a ast::ExprStringLiteral> for StringLike<'a> {
Expand Down Expand Up @@ -488,6 +496,19 @@ impl<'a> TryFrom<&'a Expr> for StringLike<'a> {
}
}

impl<'a> TryFrom<AnyNodeRef<'a>> for StringLike<'a> {
type Error = ();

fn try_from(value: AnyNodeRef<'a>) -> Result<Self, Self::Error> {
match value {
AnyNodeRef::ExprStringLiteral(value) => Ok(Self::String(value)),
AnyNodeRef::ExprBytesLiteral(value) => Ok(Self::Bytes(value)),
AnyNodeRef::ExprFString(value) => Ok(Self::FString(value)),
_ => Err(()),
}
}
}

impl Ranged for StringLike<'_> {
fn range(&self) -> TextRange {
match self {
Expand Down Expand Up @@ -561,6 +582,12 @@ impl<'a> From<&'a ast::FString> for StringLikePart<'a> {

impl<'a> From<&StringLikePart<'a>> for AnyNodeRef<'a> {
fn from(value: &StringLikePart<'a>) -> Self {
AnyNodeRef::from(*value)
}
}

impl<'a> From<StringLikePart<'a>> for AnyNodeRef<'a> {
fn from(value: StringLikePart<'a>) -> Self {
match value {
StringLikePart::String(part) => AnyNodeRef::StringLiteral(part),
StringLikePart::Bytes(part) => AnyNodeRef::BytesLiteral(part),
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -291,3 +291,83 @@
f"testeeeeeeeeeeeeeeeeeeeeeeeee{a
=}" "moreeeeeeeeeeeeeeeeeetest" # comment
)


# Trailing last-part comments

a = (
"a"
"b" # belongs to `b`
)

a: Literal[str] = (
"a"
"b" # belongs to `b`
)

a += (
"a"
"b" # belongs to `b`
)

a = (
r"a"
r"b" # belongs to `b`
)

a = (
"a"
"b"
) # belongs to the assignment

a = (((
"a"
"b" # belongs to `b`
)))

a = (((
"a"
"b"
) # belongs to the f-string expression
))

a = (
"a" "b" # belongs to the f-string expression
)

a = (
"a" "b"
# belongs to the f-string expression
)

a = (
"a"
"b" "c" # belongs to the f-string expression
)
Comment on lines +343 to +346
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I understand the change but I want to confirm as to why does this comment belongs to the f-string expression. I think the reason is that without that, the formatting might move the last part of the string to it's own line along with the comment which means the comment won't be applicable for the last second part. Is that correct?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is an edge case where there's no "right" solution because the formatter can't tell if the comment belongs to the middle part, the last part, or the entire f-string.

That's why I decided to associate it with the f-string. This matches the default behavior where we associate a comment with the outer most node that ends right before it. We could also decide to associate it with the last part, if we think this is more appropriate.


logger.error(
f"Failed to run task {task} for job"
f"with id {str(job.id)}" # type: ignore[union-attr]
)

a = (10 +
"Exception in {call_back_name} when handling msg on "
f"'{msg.topic}': '{msg.payload}'" # belongs to binary operation
)

a = 10 + (
"Exception in {call_back_name} when handling msg on "
f"'{msg.topic}': '{msg.payload}'" # belongs to last-part
)
Comment on lines +348 to +361
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It might be useful to shorten these strings (similar to other test cases) just to make sure that the formatter isn't going to collapse them because I think otherwise the line length won't even allow this to be tested.


self._attr_unique_id = (
f"{self._device.temperature.group_address_state}_"
f"{self._device.target_temperature.group_address_state}_"
f"{self._device.target_temperature.group_address}_"
f"{self._device._setpoint_shift.group_address}" # noqa: SLF001
)

return (
f"Exception in {call_back_name} when handling msg on "
f"'{msg.topic}': '{msg.payload}'" # type: ignore[str-bytes-safe]
)
110 changes: 107 additions & 3 deletions crates/ruff_python_formatter/src/comments/placement.rs
Original file line number Diff line number Diff line change
@@ -1,19 +1,19 @@
use std::cmp::Ordering;

use ast::helpers::comment_indentation_after;
use ruff_python_ast::whitespace::indentation;
use ruff_python_ast::{
self as ast, AnyNodeRef, Comprehension, Expr, ModModule, Parameter, Parameters,
self as ast, AnyNodeRef, Comprehension, Expr, ModModule, Parameter, Parameters, StringLike,
};
use ruff_python_trivia::{
find_only_token_in_range, first_non_trivia_token, indentation_at_offset, BackwardsTokenizer,
CommentRanges, SimpleToken, SimpleTokenKind, SimpleTokenizer,
};
use ruff_source_file::LineRanges;
use ruff_text_size::{Ranged, TextLen, TextRange};
use std::cmp::Ordering;

use crate::comments::visitor::{CommentPlacement, DecoratedComment};
use crate::expression::expr_slice::{assign_comment_in_slice, ExprSliceCommentSection};
use crate::expression::parentheses::is_expression_parenthesized;
use crate::other::parameters::{
assign_argument_separator_comment_placement, find_parameter_separators,
};
Expand Down Expand Up @@ -355,6 +355,41 @@ fn handle_enclosed_comment<'a>(
AnyNodeRef::ExprGenerator(generator) if generator.parenthesized => {
handle_bracketed_end_of_line_comment(comment, source)
}
AnyNodeRef::StmtReturn(_) => {
handle_trailing_implicit_concatenated_string_comment(comment, comment_ranges, source)
Copy link
Member

@dhruvmanila dhruvmanila Jan 10, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do we need to cover other statements like raise and assert?

raise Exception(
    "a"
    "b"  # belongs to `b`
)

assert False, (
    "a"
    "b"  # belongs to `b`
)

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No, only statements that use FormatStatementsLastExpression

}
AnyNodeRef::StmtAssign(assignment)
if comment.preceding_node().is_some_and(|preceding| {
preceding.ptr_eq(AnyNodeRef::from(&*assignment.value))
}) =>
{
handle_trailing_implicit_concatenated_string_comment(comment, comment_ranges, source)
}
AnyNodeRef::StmtAnnAssign(assignment)
if comment.preceding_node().is_some_and(|preceding| {
assignment
.value
.as_deref()
.is_some_and(|value| preceding.ptr_eq(value.into()))
}) =>
{
handle_trailing_implicit_concatenated_string_comment(comment, comment_ranges, source)
}
AnyNodeRef::StmtAugAssign(assignment)
if comment.preceding_node().is_some_and(|preceding| {
preceding.ptr_eq(AnyNodeRef::from(&*assignment.value))
}) =>
{
handle_trailing_implicit_concatenated_string_comment(comment, comment_ranges, source)
}
AnyNodeRef::StmtTypeAlias(assignment)
if comment.preceding_node().is_some_and(|preceding| {
preceding.ptr_eq(AnyNodeRef::from(&*assignment.value))
}) =>
{
handle_trailing_implicit_concatenated_string_comment(comment, comment_ranges, source)
}

_ => CommentPlacement::Default(comment),
}
}
Expand Down Expand Up @@ -2086,6 +2121,75 @@ fn handle_comprehension_comment<'a>(
CommentPlacement::Default(comment)
}

/// Handle end-of-line comments for parenthesized implicitly concatenated strings when used in
/// a `FormatStatementLastExpression` context:
///
/// ```python
/// a = (
/// "a"
/// "b"
/// "c" # comment
/// )
/// ```
///
/// `# comment` is a trailing comment of the last part and not a trailing comment of the entire f-string.
/// Associating the comment with the last part is important or the assignment formatting might move
/// the comment at the end of the assignment, making it impossible to suppress an error for the last part.
///
/// On the other hand, `# comment` is a trailing end-of-line f-string comment for:
///
/// ```python
/// a = (
/// "a" "b" "c" # comment
/// )
///
/// a = (
/// "a"
/// "b"
/// "c"
/// ) # comment
/// ```
///
/// Associating the comment with the f-string is desired in those cases because it allows
/// joining the string literals into a single string literal if it fits on the line.
fn handle_trailing_implicit_concatenated_string_comment<'a>(
comment: DecoratedComment<'a>,
comment_ranges: &CommentRanges,
source: &str,
) -> CommentPlacement<'a> {
if !comment.line_position().is_end_of_line() {
return CommentPlacement::Default(comment);
}

let Some(string_like) = comment
.preceding_node()
.and_then(|preceding| StringLike::try_from(preceding).ok())
else {
return CommentPlacement::Default(comment);
};

let mut parts = string_like.parts();

let (Some(last), Some(second_last)) = (parts.next_back(), parts.next_back()) else {
return CommentPlacement::Default(comment);
};

if source.contains_line_break(TextRange::new(second_last.end(), last.start()))
&& is_expression_parenthesized(string_like.as_expression_ref(), comment_ranges, source)
{
let range = TextRange::new(last.end(), comment.start());

if !SimpleTokenizer::new(source, range)
.skip_trivia()
.any(|token| token.kind() == SimpleTokenKind::RParen)
{
return CommentPlacement::trailing(AnyNodeRef::from(last), comment);
}
}

CommentPlacement::Default(comment)
}

/// Returns `true` if the parameters are parenthesized (as in a function definition), or `false` if
/// not (as in a lambda).
fn are_parameters_parenthesized(parameters: &Parameters, contents: &str) -> bool {
Expand Down
2 changes: 1 addition & 1 deletion crates/ruff_python_formatter/src/string/implicit.rs
Original file line number Diff line number Diff line change
Expand Up @@ -99,7 +99,7 @@ impl Format<PyFormatContext<'_>> for FormatImplicitConcatenatedStringExpanded<'_
StringLikePart::FString(part) => part.format().fmt(f),
});

let part_comments = comments.leading_dangling_trailing(&part);
let part_comments = comments.leading_dangling_trailing(part);
joiner.entry(&format_args![
leading_comments(part_comments.leading),
format_part,
Expand Down
Original file line number Diff line number Diff line change
@@ -1,7 +1,6 @@
---
source: crates/ruff_python_formatter/tests/fixtures.rs
input_file: crates/ruff_python_formatter/resources/test/fixtures/ruff/expression/fstring.py
snapshot_kind: text
---
## Input
```python
Expand Down
Loading
Loading