[stdlib] Allow a default for optional interpolations #80547

natecook1000 · 2025-04-04T20:22:44Z

This adds an appendInterpolation overload to DefaultStringInterpolation that includes a parameter for providing a default string when the value to interpolate is nil. This allows this kind of usage:

let age: Int? = nil
print("Your age is \(age, default: "timeless")")
// Prints "Your age is timeless"

Includes an additional fixit when optional values are interpolated to use this default: parameter.

This adds an `appendInterpolation` overload to `DefaultStringInterpolation` that includes a parameter for providing a default string when the value to interpolate is `nil`. This allows this kind of usage: let age: Int? = nil print("Your age is \(age, default: "timeless")") // Prints "Your age is timeless" Includes an additional fixit when optional values are interpolated to use this `default:` parameter.

natecook1000 · 2025-04-04T20:37:47Z

lib/Sema/MiscDiagnostics.cpp

@@ -5688,6 +5688,11 @@ static void diagnoseUnintendedOptionalBehavior(const Expr *E,
          .fixItInsertAfter(arg->getEndLoc(), ")");

      if (kind == UnintendedInterpolationKind::Optional) {
+        // Suggest using a default interpolation value parameter.
+        Ctx.Diags.diagnose(arg->getLoc(), diag::default_optional_parameter)


@beccadax Do you know if it would be possible to only provide this fixit when the string interpolation type is DefaultStringInterpolation?

If you had the caller extract the ConcreteDeclRef and pass it in, then yes, that ought to be possible.

As an aside, I'm thinking we might not want to suggest both \(_:default:) and ??, as they mean nearly the same thing. Maybe ?? if it's an Optional<String> and \(_:default:) otherwise?

I updated the logic so that the new fixit is only provided for DefaultStringInterpolation types, and isn't provided for optional strings, since it's the same functionality as the preferred ??. I left in the ?? suggestion for other optional types, since there could be some types or situations where a default value of that same type still makes sense.

(I also moved the String(describing:) fix-it to be the last suggestion, since it feels like "fix this warning" should be preferred over "silence this warning".)

beccadax · 2025-04-08T23:53:10Z

lib/Sema/MiscDiagnostics.cpp

@@ -5688,6 +5688,11 @@ static void diagnoseUnintendedOptionalBehavior(const Expr *E,
          .fixItInsertAfter(arg->getEndLoc(), ")");

      if (kind == UnintendedInterpolationKind::Optional) {
+        // Suggest using a default interpolation value parameter.
+        Ctx.Diags.diagnose(arg->getLoc(), diag::default_optional_parameter)


If you had the caller extract the ConcreteDeclRef and pass it in, then yes, that ought to be possible.

As an aside, I'm thinking we might not want to suggest both \(_:default:) and ??, as they mean nearly the same thing. Maybe ?? if it's an Optional<String> and \(_:default:) otherwise?

beccadax · 2025-04-08T23:54:48Z

stdlib/public/core/StringInterpolation.swift

+  }
+
+  @backDeployed(before: SwiftStdlib 6.2)
+  @_disfavoredOverload


Instead of using @_disfavoredOverload for these, you should be able to provide a fourth T: TextOutputStreamable, T: CustomStringConvertible overload. That's how regular appendInterpolation(_:) works. (The type checker really penalizes disfavored overloads, so it's usually better to avoid it if you can.)

Also, consider making these @inlinable. When I originally implemented DefaultStringInterpolation, I did some interpolation performance benchmarking (the test should still be in the benchmark suite) and found that inlining really matters for it.

Oh, I didn't realize! I thought I was helping out by only providing three overloads instead of four. 🙃 I'll look at adding @inlinable, too. Thanks!

This fix-it suppresses the string interpolation warning, but doesn't provide any additional affordance for dealing with the problem, so it shouldn't be the first suggestion.

Modify fix-its so the new interpolation API is only recommended for non-strings (since `??` is a more fluent resolution for strings) and when used in a `DefaultStringInterpolation` type (since other interpolation types won't have this API).

xwu · 2025-04-15T22:06:03Z

@swift-ci Please smoke test

natecook1000 · 2025-04-16T01:24:56Z

@swift-ci Please smoke test

natecook1000 · 2025-04-18T17:09:30Z

@swift-ci Please smoke test

natecook1000 · 2025-04-18T21:28:01Z

@swift-ci Please smoke test

xwu · 2025-04-19T14:23:25Z

@swift-ci Please build toolchain

xwu · 2025-04-21T18:08:41Z

@swift-ci Please build macOS toolchain

natecook1000 · 2025-04-21T19:48:32Z

@swift-ci Please build toolchain macOS platform

natecook1000 · 2025-04-30T16:06:40Z

@swift-ci Please smoke test

natecook1000 · 2025-04-30T19:30:17Z

@swift-ci Please smoke test

natecook1000 · 2025-04-30T23:36:31Z

@swift-ci Please smoke test Linux platform

natecook1000 · 2025-05-01T06:01:35Z

@swift-ci Please smoke test Linux platform

amartini51

Looks good — left minor feedback and a suggestion.

amartini51 · 2025-05-01T20:44:56Z

stdlib/public/core/StringInterpolation.swift

+  /// Interpolates the given optional value's textual representation, or the
+  /// specified default string, into the string literal being created.
+  ///
+  /// You don't need to call this method directly. It is used by the compiler


Here and in the other places this phrasing appears, missing a contraction:

Suggested change

/// You don't need to call this method directly. It is used by the compiler

/// You don't need to call this method directly. It's used by the compiler

amartini51 · 2025-05-01T20:57:33Z

stdlib/public/core/StringInterpolation.swift

+  ///
+  /// You don't need to call this method directly. It is used by the compiler
+  /// when interpreting string interpolations where you provide a `default`
+  /// parameter.


The code listing would benefit from a sentence introducing it — that might also give you an earlier place to frame the fact that the default interpolation value is used when the value's nil.

amartini51 · 2025-05-02T22:02:57Z

Changes in 3973a36 look good to me!

natecook1000 · 2025-05-05T16:50:10Z

Now that this has evolution approval, it's ready for review – cc @beccadax @Azoy

Azoy · 2025-05-05T17:00:00Z

stdlib/public/core/StringInterpolation.swift

+  ///   - value: The value to include in a string interpolation, if non-`nil`.
+  ///   - default: The string to include if `value` is `nil`.
+  @backDeployed(before: SwiftStdlib 6.2)
+  @inlinable


Any reason for @backDeployed+@inlinable instead of just @_alwaysEmitIntoClient?

It seemed like the right set, presuming that @inlinable has different behavior than @_alwaysEmit.... Do we have guidance about which to prefer/when between inlinable and always-emit?

These are very simple functions and I'm unsure if we want to commit these as ABI. With AEIC, the stdlib doesn't take the ABI hit and clients who don't use it don't pay for it either.

natecook1000 · 2025-05-06T20:09:31Z

@swift-ci Please smoke test

natecook1000 · 2025-05-07T04:29:28Z

@swift-ci Please smoke test Windows platform

This adds an `appendInterpolation` overload to `DefaultStringInterpolation` that includes a parameter for providing a default string when the value to interpolate is `nil`. This allows this kind of usage: ```swift let age: Int? = nil print("Your age is \(age, default: "timeless")") // Prints "Your age is timeless" ``` The change includes an additional fixit when optional values are interpolated, with a suggestion to use this `default:` parameter.

natecook1000 added standard library Area: Standard library umbrella swift evolution proposal needed Flag → feature: A feature that warrants a Swift evolution proposal labels Apr 4, 2025

natecook1000 commented Apr 4, 2025

View reviewed changes

beccadax reviewed Apr 8, 2025

View reviewed changes

natecook1000 added 3 commits April 13, 2025 17:49

Merge branch 'main' into string-interp-default

cb27a9e

Move String(describing:) fix to be last

5c197d3

This fix-it suppresses the string interpolation warning, but doesn't provide any additional affordance for dealing with the problem, so it shouldn't be the first suggestion.

Limit when \(_:default:) is recommended

0f80542

Modify fix-its so the new interpolation API is only recommended for non-strings (since `??` is a more fluent resolution for strings) and when used in a `DefaultStringInterpolation` type (since other interpolation types won't have this API).

Add proper overload for string interpolation

476a0ec

natecook1000 added 2 commits April 18, 2025 12:08

Update IDE completion tests

652dbcc

Update other tests with new APIs

5535589

natecook1000 requested a review from beccadax April 18, 2025 17:11

One more slight test update

1d50b16

natecook1000 added 3 commits April 30, 2025 11:04

Let default parameter take some StringProtocol

d4893dc

Add documentation

c06a8d2

Relax related wording.

2a0598b

natecook1000 added 2 commits April 30, 2025 14:28

Merge branch 'main' into string-interp-default

5e295c1

Update api-digester test for new format

0120db6

natecook1000 removed the swift evolution proposal needed Flag → feature: A feature that warrants a Swift evolution proposal label May 1, 2025

natecook1000 added the swift evolution approved Flag → feature: A feature that was approved through the Swift evolution process label May 1, 2025

natecook1000 marked this pull request as ready for review May 1, 2025 06:02

natecook1000 requested review from a team, hborla, slavapestov, xedin, ahoppen, bnbarham, hamishknight and rintaro as code owners May 1, 2025 06:02

natecook1000 requested a review from amartini51 May 1, 2025 13:34

natecook1000 changed the title ~~[DRAFT] Allow a default for optional interpolations~~ [stdlib] Allow a default for optional interpolations May 1, 2025

amartini51 approved these changes May 1, 2025

View reviewed changes

Improve string interpolation docs

3973a36

Azoy reviewed May 5, 2025

View reviewed changes

natecook1000 added 3 commits May 6, 2025 13:19

Switch to always-emit

02cdb70

Merge branch 'main' into string-interp-default

264a1ba

Update ABI/API test fixtures

981d88b

natecook1000 merged commit e68069f into swiftlang:main May 7, 2025
3 checks passed

natecook1000 deleted the string-interp-default branch May 7, 2025 17:47

natecook1000 mentioned this pull request May 7, 2025

[6.2][stdlib] Allow a default for optional interpolations #81360

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[stdlib] Allow a default for optional interpolations #80547

[stdlib] Allow a default for optional interpolations #80547

natecook1000 commented Apr 4, 2025

natecook1000 Apr 4, 2025

beccadax Apr 8, 2025

natecook1000 Apr 18, 2025

beccadax Apr 8, 2025

beccadax Apr 8, 2025

natecook1000 Apr 10, 2025

xwu commented Apr 15, 2025

natecook1000 commented Apr 16, 2025

natecook1000 commented Apr 18, 2025

natecook1000 commented Apr 18, 2025

xwu commented Apr 19, 2025

xwu commented Apr 21, 2025

natecook1000 commented Apr 21, 2025

natecook1000 commented Apr 30, 2025

natecook1000 commented Apr 30, 2025

natecook1000 commented Apr 30, 2025

natecook1000 commented May 1, 2025

amartini51 left a comment

amartini51 May 1, 2025

amartini51 May 1, 2025

amartini51 commented May 2, 2025

natecook1000 commented May 5, 2025

Azoy May 5, 2025

natecook1000 May 5, 2025

Azoy May 6, 2025 •

edited

Loading

natecook1000 commented May 6, 2025

natecook1000 commented May 7, 2025

	/// You don't need to call this method directly. It is used by the compiler
	/// You don't need to call this method directly. It's used by the compiler

[stdlib] Allow a default for optional interpolations #80547

[stdlib] Allow a default for optional interpolations #80547

Conversation

natecook1000 commented Apr 4, 2025

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

xwu commented Apr 15, 2025

natecook1000 commented Apr 16, 2025

natecook1000 commented Apr 18, 2025

natecook1000 commented Apr 18, 2025

xwu commented Apr 19, 2025

xwu commented Apr 21, 2025

natecook1000 commented Apr 21, 2025

natecook1000 commented Apr 30, 2025

natecook1000 commented Apr 30, 2025

natecook1000 commented Apr 30, 2025

natecook1000 commented May 1, 2025

amartini51 left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

amartini51 commented May 2, 2025

natecook1000 commented May 5, 2025

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Azoy May 6, 2025 • edited Loading

Choose a reason for hiding this comment

natecook1000 commented May 6, 2025

natecook1000 commented May 7, 2025

Azoy May 6, 2025 •

edited

Loading