Automated Resyntax fixes by resyntax-ci[bot] · Pull Request #1438 · racket/typed-racket

resyntax-ci · 2025-02-28T00:35:13Z

Resyntax fixed 41 issues in 20 files.

Fixed 15 occurrences of single-clause-match-to-match-define
Fixed 6 occurrences of let-to-define
Fixed 2 occurrences of define-simple-macro-to-define-syntax-parse-rule
Fixed 2 occurrences of if-else-false-to-and
Fixed 2 occurrences of zero-comparison-to-positive?
Fixed 2 occurrences of map-to-for
Fixed 2 occurrences of sort-with-keyed-comparator-to-sort-by-key
Fixed 1 occurrence of case-lambda-with-single-case-to-lambda
Fixed 1 occurrence of if-begin-to-cond
Fixed 1 occurrence of define-lambda-to-define
Fixed 1 occurrence of inverted-unless
Fixed 1 occurrence of provide-deduplication
Fixed 1 occurrence of begin0-let-to-define-begin0
Fixed 1 occurrence of inverted-when
Fixed 1 occurrence of define-values-values-to-define
Fixed 1 occurrence of cond-let-to-cond-define
Fixed 1 occurrence of syntax-disarm-migration

The `define-simple-macro` form has been renamed to `define-syntax-parse-rule`.

This expression is equivalent to calling the `positive?` predicate.

The `syntax-disarm` function is a legacy function that does nothing.

Internal definitions are recommended instead of `let` expressions, to reduce nesting.

This `if` expression can be refactored to an equivalent expression using `and`.

This `match` expression can be simplified using `match-define`.

This `map` operation can be replaced with a `for/list` loop.

The `define` form supports a shorthand for defining functions.

Providing the same identifier multiple times is unnecessary.

Using `cond` instead of `if` here makes `begin` unnecessary

The `let` expression in this `begin0` form can be extracted into the surrounding definition context.

This `case-lambda` form only has one case. Use a regular lambda instead.

This use of `define-values` is unnecessary.

This negated `unless` expression can be replaced by a `when` expression.

This `sort` expression can be replaced with a simpler, equivalent expression.

This negated `when` expression can be replaced by an `unless` expression.

Internal definitions are recommended instead of `let` expressions, to reduce nesting.

samth · 2025-03-04T17:28:33Z

typed-racket-lib/typed-racket/core.rkt

-                                              (when (not (eq? te-mode deep))
+                                              (unless (eq? te-mode deep)
                                                (raise-arguments-error
-                                                  (string->symbol (format "typed/racket/~a" (keyword->string (syntax-e te-attr))))


@jackfirth this could use format-symbol from racket/syntax

Added a new rule for this in jackfirth/resyntax#442.

samth · 2025-03-04T17:30:32Z

typed-racket-lib/typed-racket/rep/prop-rep.rkt

  [#:for-each (f) (for-each f ps)]
  [#:custom-constructor/contract
   (-> (listof (or/c TypeProp? NotTypeProp? LeqProp?)) OrProp?)
-   (let ([ps (sort ps (λ (p q) (unsafe-fx<= (eq-hash-code p)


@jackfirth my guess is that this change is slower

If it is, that seems like a missed optimization opportunity in sort.

It's really an issue of inlining.

samth · 2025-03-04T17:30:56Z

typed-racket-lib/typed-racket/static-contracts/combinators/control.rkt

-      (map f* vals)
-      (and call-cc (map f* call-cc)))]))
+  (match-define (pt-seq vals call-cc) seq)
+  (define (f* a)


@sorawee why is this on two lines?

It's a "community"'s preference to always format function definition this way. It used to support one-line format, but IIRC, either @jackfirth or @Metaxal gathered supporters to NOT do the one-line format, and I made the change accordingly.

I would be interested to find that discussion; I strongly disagree.

If it was me, I don't remember having a strong opinion about this. Perhaps it was more of a pragmatic choice at that time(?). But again, perhaps it wasn't me 😁

It's @Metaxal. I'm very lucky that I wrote your name in the commit message 😂

sorawee/fmt@831bb4c

Haha indeed! Though "Suggestion by" sounds a bit different from "gathered supporters to NOT do" ;)

Again, I don't feel particularly strongly about this, although I do have a preference for two lines so that the function header stands out. I'm not going to die on this hill though.

It was also me. I was (and still am) of the strong opinion that function definitions should never be on a single line. Otherwise it's hard to tell them apart from variable definitions.

Functions are just another kind of variable. ;)

samth · 2025-03-04T17:32:57Z

typed-racket-lib/typed-racket/typed-reader.rkt

+        (let-values ([(l c p) (port-next-location port)])
+          (raise-read-error (format "typed expression ~a not properly terminated"
+                                    (syntax->datum name))
+                            src


I hate this indentation.

What's your preferred format?

having all the additional arguments on the same line. I realize that this preference of mine is not widely shared but I'm still grumpy.

FWIW, I also find the new indentation jarring: there's a big empty space with just a few letters. It's quite wasteful information-wise.

I agree with this particular case, but I'm not sure how to turn this into code in a way that makes other cases also look good.

E.g.: compare:

(format "This is a very long string" first-long-string second-long-string third-long-string fourth-long-string))

vs:

(format "This is a very long string" first-long-string second-long-string third-long-string fourth-long-string))

I think the second one is preferable.

Yes, I agree with that. I bet you can make resyntax do it though. :)

I could attempt something. Do you think this case with raise-read-error comes up enough to be worth it?

No it would have to be more general, but I bet the case where format is involved happens a lot.

More general is trickier because it makes it difficult to pick a good name for the introduced variable.

Good excuse to use LLM to find a name?

samth · 2025-04-18T14:21:33Z

Should this be merged or closed?

jackfirth · 2025-04-18T19:31:38Z

@samth Up to you. Resyntax cannot (yet) update pull requests after it's made them so the options are to merge it as-is, merge it with manual edits for anything you want to change, or close it and let Resyntax try again next time. As the repo owner I defer to you on that choice.

samth · 2025-04-18T19:34:23Z

Ok I'm going to close this and we'll see what the next one does.

resyntax-ci bot added 18 commits February 28, 2025 00:35

Fix 2 occurrences of define-simple-macro-to-define-syntax-parse-rule

21784fd

The `define-simple-macro` form has been renamed to `define-syntax-parse-rule`.

Fix 2 occurrences of zero-comparison-to-positive?

63a7fe1

This expression is equivalent to calling the `positive?` predicate.

Fix 1 occurrence of syntax-disarm-migration

60a1399

The `syntax-disarm` function is a legacy function that does nothing.

Fix 5 occurrences of let-to-define

da8c553

Internal definitions are recommended instead of `let` expressions, to reduce nesting.

Fix 2 occurrences of if-else-false-to-and

583a1e3

This `if` expression can be refactored to an equivalent expression using `and`.

Fix 15 occurrences of single-clause-match-to-match-define

f8d6c9d

This `match` expression can be simplified using `match-define`.

Fix 2 occurrences of map-to-for

84dc345

This `map` operation can be replaced with a `for/list` loop.

Fix 1 occurrence of define-lambda-to-define

c6f5a28

The `define` form supports a shorthand for defining functions.

Fix 1 occurrence of provide-deduplication

5f1257e

Providing the same identifier multiple times is unnecessary.

Fix 1 occurrence of if-begin-to-cond

57862b3

Using `cond` instead of `if` here makes `begin` unnecessary

Fix 1 occurrence of begin0-let-to-define-begin0

651bd68

The `let` expression in this `begin0` form can be extracted into the surrounding definition context.

Fix 1 occurrence of case-lambda-with-single-case-to-lambda

3575e85

This `case-lambda` form only has one case. Use a regular lambda instead.

Fix 1 occurrence of define-values-values-to-define

3f21958

This use of `define-values` is unnecessary.

Fix 1 occurrence of inverted-unless

f2601c5

This negated `unless` expression can be replaced by a `when` expression.

Fix 2 occurrences of sort-with-keyed-comparator-to-sort-by-key

c7858d9

This `sort` expression can be replaced with a simpler, equivalent expression.

Fix 1 occurrence of inverted-when

c9d12e3

This negated `when` expression can be replaced by an `unless` expression.

Fix 1 occurrence of let-to-define

a51e0f3

Internal definitions are recommended instead of `let` expressions, to reduce nesting.

Fix 1 occurrence of cond-let-to-cond-define

df62579

Internal definitions are recommended instead of `let` expressions, to reduce nesting.

samth reviewed Mar 4, 2025

View reviewed changes

jackfirth mentioned this pull request Mar 9, 2025

Add format-string-to-format-symbol rule jackfirth/resyntax#442

Merged

samth closed this Apr 18, 2025

Uh oh!

Conversation

resyntax-ci bot commented Feb 28, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

samth commented Apr 18, 2025

Uh oh!

jackfirth commented Apr 18, 2025

Uh oh!

samth commented Apr 18, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants