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

Pass parent to NeedsParentheses #5708

Merged
merged 1 commit into from
Jul 13, 2023
Merged

Conversation

MichaReiser
Copy link
Member

@MichaReiser MichaReiser commented Jul 12, 2023

Summary

This PR refactors NeedsParentheses so that implementations get access to the parent node. Having access to the parent node is e.g. required for omitting parentheses from the walrus operator when not strictly necessary.

The main challenge that I ran into is that simply storing the parent node in Parenthesize::Optional isn't possible because Parenthesize then requires lifetimes, and AsFormat (with FormatRuleWithOptions) does not support lifetimes... ugh

This forced me to rethink the problem and I'm not entirely unhappy with the outcome (There's room for improvement on the naming).

  • FormatExpr: Accepts a Parentheses option that can either be Always, Never, Preserve, where Preserve is the default. This covers the base case of expression formatting
  • maybe_parenthesize_expression is a new builder that may parenthesize an expression depending on Parenthesized and the expressions NeedsParentheses implementation.

I went along and removed Parentheses::Custom as part of this PR. This caused me some headaches with string formatting until I realized that we currently make an exception for string formatting which we do not for other nodes. The old implementation forced strings to be flat when used in an Expression statement to match Black, but we don't enforce the same for binary expressions and other expressions nodes. My take is that we should solve this holistically rather than for some nodes only.

Test Plan

Most of this is pure refactor. I reviewed the snapshot tests and they seem reasonable.

@MichaReiser
Copy link
Member Author

MichaReiser commented Jul 12, 2023

Current dependencies on/for this PR:

This comment was auto-generated by Graphite.

+ )(db.not_(User.is_pending.astext.cast(db.Boolean).is_(True))).order_by(
+ User.created_at.desc()
+ ).with_for_update(key_share=True).all()
+ filter(db.not_(User.is_pending.astext.cast(db.Boolean).is_(True))).order_by(
Copy link
Member Author

Choose a reason for hiding this comment

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

Requires call chain formatting

Copy link
Member

Choose a reason for hiding this comment

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

yep i had opened #5343 to track this

for (
z
) in (NOT_YET_IMPLEMENTED_generator_key for NOT_YET_IMPLEMENTED_generator_key in []):
for z in (NOT_YET_IMPLEMENTED_generator_key for NOT_YET_IMPLEMENTED_generator_key in []):
Copy link
Member Author

Choose a reason for hiding this comment

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

Requires proper generator formatting

@@ -203,7 +203,20 @@ String \"\"\"

"Let's" "start" "with" "a" "simple" "example"

"Let's" "start" "with" "a" "simple" "example" "now repeat after me:" "I am confident" "I am confident" "I am confident" "I am confident" "I am confident"
(
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 per se is a regression because black keeps the string flat inside of ExprStmt. However, this is the same as that black does not try to wrap binary expressions on statement level (except if they contain a list). I decided to remove the work around for strings so that we have the same behavior for all expressions and either solve it holistically or not at all.

Copy link
Member

Choose a reason for hiding this comment

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

implicit string concatenation (vs. triple quoted strings) at statement level seems like an odd choice, do you by chance have any usage examples?

Copy link
Member Author

Choose a reason for hiding this comment

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

I didn't find any in the whole django project... (as expected?)

_parent: AnyNodeRef,
_context: &PyFormatContext,
) -> OptionalParentheses {
OptionalParentheses::Never
Copy link
Member Author

Choose a reason for hiding this comment

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

Names never require optional parentheses... There's nothing to split by

Parentheses::Optional => Parentheses::Never,
parentheses => parentheses,
) -> OptionalParentheses {
self.func.needs_parentheses(parent, context)
Copy link
Member Author

Choose a reason for hiding this comment

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

Never is incorrect. It requires parentheses if the func node requires parentheses (see attribute chain)

Copy link
Member

Choose a reason for hiding this comment

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

yep i should have placed a todo comment about missing fluent style support there

@MichaReiser MichaReiser changed the base branch from main to right-parens-join-comma July 12, 2023 13:42
@MichaReiser MichaReiser requested a review from konstin July 12, 2023 13:51
@MichaReiser MichaReiser marked this pull request as ready for review July 12, 2023 13:51
This was referenced Jul 12, 2023
if self.value.is_str() {
let contents = context.locator().slice(self.range());
// Don't wrap triple quoted strings
if is_multiline_string(self, context.source()) || !is_implicit_concatenation(contents) {
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 somewhat slow. We should consider extracting the ranges of all multiline strings (similar to Ruff)

@MichaReiser MichaReiser added the formatter Related to the formatter label Jul 12, 2023
@github-actions
Copy link
Contributor

github-actions bot commented Jul 12, 2023

PR Check Results

Ecosystem

✅ ecosystem check detected no changes.

Benchmark

Linux

group                                      main                                   pr
-----                                      ----                                   --
formatter/large/dataset.py                 1.01      8.0±0.02ms     5.1 MB/sec    1.00      7.9±0.02ms     5.2 MB/sec
formatter/numpy/ctypeslib.py               1.02   1871.4±2.93µs     8.9 MB/sec    1.00   1832.7±2.17µs     9.1 MB/sec
formatter/numpy/globals.py                 1.00    204.5±0.38µs    14.4 MB/sec    1.00    204.8±0.52µs    14.4 MB/sec
formatter/pydantic/types.py                1.01      4.0±0.01ms     6.3 MB/sec    1.00      4.0±0.01ms     6.4 MB/sec
linter/all-rules/large/dataset.py          1.00     13.3±0.02ms     3.0 MB/sec    1.01     13.4±0.03ms     3.0 MB/sec
linter/all-rules/numpy/ctypeslib.py        1.00      3.4±0.01ms     4.9 MB/sec    1.00      3.4±0.02ms     4.9 MB/sec
linter/all-rules/numpy/globals.py          1.00    431.2±0.58µs     6.8 MB/sec    1.00    431.3±0.46µs     6.8 MB/sec
linter/all-rules/pydantic/types.py         1.05      6.3±0.08ms     4.1 MB/sec    1.00      6.0±0.02ms     4.3 MB/sec
linter/default-rules/large/dataset.py      1.00      6.7±0.01ms     6.1 MB/sec    1.01      6.8±0.02ms     6.0 MB/sec
linter/default-rules/numpy/ctypeslib.py    1.00   1487.4±1.40µs    11.2 MB/sec    1.00   1482.8±1.61µs    11.2 MB/sec
linter/default-rules/numpy/globals.py      1.00    169.2±0.86µs    17.4 MB/sec    1.00    168.9±0.22µs    17.5 MB/sec
linter/default-rules/pydantic/types.py     1.00      3.1±0.01ms     8.3 MB/sec    1.00      3.1±0.01ms     8.4 MB/sec

Windows

group                                      main                                   pr
-----                                      ----                                   --
formatter/large/dataset.py                 1.01     12.4±0.59ms     3.3 MB/sec    1.00     12.2±0.49ms     3.3 MB/sec
formatter/numpy/ctypeslib.py               1.00      2.7±0.13ms     6.1 MB/sec    1.01      2.8±0.15ms     6.0 MB/sec
formatter/numpy/globals.py                 1.00   305.5±18.11µs     9.7 MB/sec    1.04   317.8±22.72µs     9.3 MB/sec
formatter/pydantic/types.py                1.00      6.0±0.27ms     4.3 MB/sec    1.01      6.0±0.32ms     4.2 MB/sec
linter/all-rules/large/dataset.py          1.02     20.9±0.92ms  1993.5 KB/sec    1.00     20.6±0.70ms  2024.7 KB/sec
linter/all-rules/numpy/ctypeslib.py        1.00      5.3±0.22ms     3.2 MB/sec    1.04      5.5±0.52ms     3.0 MB/sec
linter/all-rules/numpy/globals.py          1.00   643.5±39.84µs     4.6 MB/sec    1.00   644.1±26.30µs     4.6 MB/sec
linter/all-rules/pydantic/types.py         1.01      9.2±0.33ms     2.8 MB/sec    1.00      9.0±0.39ms     2.8 MB/sec
linter/default-rules/large/dataset.py      1.03     10.7±0.42ms     3.8 MB/sec    1.00     10.3±0.54ms     3.9 MB/sec
linter/default-rules/numpy/ctypeslib.py    1.02      2.2±0.10ms     7.5 MB/sec    1.00      2.2±0.09ms     7.7 MB/sec
linter/default-rules/numpy/globals.py      1.00   268.5±16.21µs    11.0 MB/sec    1.01   272.3±13.31µs    10.8 MB/sec
linter/default-rules/pydantic/types.py     1.00      4.7±0.15ms     5.5 MB/sec    1.01      4.7±0.24ms     5.5 MB/sec

Parentheses::Optional => Parentheses::Never,
parentheses => parentheses,
) -> OptionalParentheses {
self.func.needs_parentheses(parent, context)
Copy link
Member

Choose a reason for hiding this comment

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

yep i should have placed a todo comment about missing fluent style support there

+ )(db.not_(User.is_pending.astext.cast(db.Boolean).is_(True))).order_by(
+ User.created_at.desc()
+ ).with_for_update(key_share=True).all()
+ filter(db.not_(User.is_pending.astext.cast(db.Boolean).is_(True))).order_by(
Copy link
Member

Choose a reason for hiding this comment

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

yep i had opened #5343 to track this

@@ -203,7 +203,20 @@ String \"\"\"

"Let's" "start" "with" "a" "simple" "example"

"Let's" "start" "with" "a" "simple" "example" "now repeat after me:" "I am confident" "I am confident" "I am confident" "I am confident" "I am confident"
(
Copy link
Member

Choose a reason for hiding this comment

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

implicit string concatenation (vs. triple quoted strings) at statement level seems like an odd choice, do you by chance have any usage examples?

_context: &PyFormatContext,
) -> OptionalParentheses {
// Unlike tuples, named expression parentheses are not part of the range even when
// mandatory. See [PEP 572](https://peps.python.org/pep-0572/) for details.
Copy link
Member

Choose a reason for hiding this comment

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

it's actually a TODO on my end

Suggested change
// mandatory. See [PEP 572](https://peps.python.org/pep-0572/) for details.
// mandatory. TODO(konstin): Implement [PEP 572](https://peps.python.org/pep-0572/) rules.

Base automatically changed from right-parens-join-comma to main July 12, 2023 16:21
@MichaReiser MichaReiser merged commit 067b2a6 into main Jul 13, 2023
16 checks passed
@MichaReiser MichaReiser deleted the refactor-needs-parentheses branch July 13, 2023 06:57
konstin pushed a commit that referenced this pull request Jul 19, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
formatter Related to the formatter
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants