[Requirement Machine] Implement same-element requirements. #67465

hborla · 2023-07-22T07:56:05Z

SE-0393 supports same-element requirements. From the proposal:

A same-type requirement where one side is a type parameter pack and the other type is a scalar type that does not capture any type parameter packs is interpreted as constraining each element of the replacement type pack to the same scalar type:
func variadic<each S: Sequence, T>(_: repeat each S) where repeat (each S).Element == T {}
This is called a same-element requirement.

We represent pack elements in the requirement machine through a new singleton PackElement symbol, written [element]. Same-element requirements are mapped to re-write rules that prefix type parameter pack terms with the [element] symbol.

For example, given a protocol P with an associated type A, the generic signature <each T, U> where repeat each T: P, repeat (each T).A == U maps to the following re-write system:

Rewrite system: {
- [P].[P] => [P]
- [P].A => [P:A]
- [P].[P:A] => [P:A]
- [P:A].[shape] => [shape]
- τ_0_1.[shape] => [shape]
- each τ_0_0.[P] => each τ_0_0
- [element].each τ_0_0.A => τ_0_1
- each τ_0_0.A => each τ_0_0.[P:A]
- [element].each τ_0_0.[P:A] => τ_0_1
- [element].each τ_0_0.[shape] => [shape]
}

The implementation is gated behind -enable-experimental-feature SameElementRequirements because the minimal conformance computation doesn't yet support same-element requirements. The current rewrite rules also do not support deriving pack element conformance from a scalar type parameter conformance through same-element requirements. For example, the generic signature <each T, U> where U: P, repeat each T == U maps to the following rewrite system:

Rewrite system: {
- [P].[P] => [P]
- [P].A => [P:A]
- [P].[P:A] => [P:A]
- [P:A].[shape] => [shape]
- τ_0_1.[shape] => [shape]
- τ_0_1.[P] => τ_0_1
- [element].each τ_0_0 => τ_0_1
- τ_0_1.A => τ_0_1.[P:A]
}

The conformance rule τ_0_1.[P] => τ_0_1 and the same-element rule [element].each τ_0_0 => τ_0_1 do not provide a rewrite path to the pack conformance rule each τ_0_0.[P] => each τ_0_0. One possible solution is to use element symbols for all property rules, which would also allow the minimal conformance support to fallout naturally. However, this approach poses a problem for same-type-pack requirements. If we use [element] symbols for same-type-pack requirements, e.g. [element].each τ_0_0 => [element].each τ_0_1, then this rule no longer implies a same-shape rule each τ_0_0.[shape] => each τ_0_1.[shape]. Stripping the [element] symbols from same-type-pack rewrite rules creates a different set of problems for conformances, because the pack conformance rules contain [element] symbols, which would prohibit applying associated type rewrite rules to non-element pack terms. I'm still exploring strategies for resolving these issues, hence the experimental flag!

lib/AST/RequirementMachine/Rule.cpp

lib/AST/RequirementMachine/RewriteSystem.cpp

slavapestov · 2023-08-31T15:58:10Z

lib/AST/RequirementMachine/RuleBuilder.cpp

+      // requirement that becomes the following rewrite rule:
+      //
+      //   [element].T.[concrete: C<X, Y>] => [element].T
+      if (subjectType->isParameterPack()) {


But only if the constraint type does not contain any packs right?

Yeah, that's right. ~~I'm not sure what the rewrite rule looks like if the constraint type contains packs but the subject type does not? That's still conceptually a same-element requirement.~~ Slava reminded me that something like repeat T == Array<each U> is invalid because it's not representable!

slavapestov · 2023-08-31T15:59:04Z

lib/AST/RequirementMachine/Symbol.cpp

@@ -80,6 +80,14 @@ struct Symbol::Storage final
    Kind = Kind::Shape;
  }

+  /// A dummy type for overload resolution of the
+  /// 'pack element' constructor for Storage.
+  struct ForPackElement {};


I think you can get rid of both ForPack and ForShape and just add a Storage(Kind) overload, but it's a matter of taste.

hborla · 2023-09-02T21:18:18Z

@swift-ci please smoke test

… elements.

…equirements.

opened element generic environments containing same-element requirements.

…mbols.

…ure flag.

hborla · 2023-09-03T00:30:16Z

@swift-ci please smoke test

slavapestov · 2023-09-03T02:23:14Z

lib/AST/GenericSignature.cpp

+        // pack element must always be on the left side of the rule, so type
+        // parameters can appear out of order in the final same-type requirement.
+        if (firstType->isParameterPack() == secondType->isParameterPack() &&
+            compareDependentTypes(firstType, secondType) >= 0) {


Can you change RequirementBuilder.cpp to build them in the other order instead?

xedin

The constraint solver changes look good to me, I left a small comment inline about locator changes.

xedin · 2023-09-04T07:05:10Z

lib/Sema/CSSimplify.cpp

@@ -13584,6 +13584,16 @@ ConstraintSystem::SolutionKind ConstraintSystem::simplifyShapeOfConstraint(
    return SolutionKind::Solved;
  }

+  if (isSingleUnlabeledPackExpansionTuple(packTy)) {
+    auto *packVar =
+        createTypeVariable(getConstraintLocator(locator), TVO_CanBindToPack);


I think the locator for everything here except to the new ShapeOf constraint needs to be locator without its last element (which is PackShape).

xedin · 2023-09-04T07:22:26Z

lib/Sema/CSGen.cpp

@@ -3125,18 +3125,9 @@ namespace {
        auto expansionType =
            CS.getType(packEnvironment)->castTo<PackExpansionType>();
        CS.addConstraint(ConstraintKind::ShapeOf, expansionType->getCountType(),
-                         elementType,
+                         packType,
                         CS.getConstraintLocator(packEnvironment,
                                                 ConstraintLocator::PackShape));


Interesting, this was completely backwards but worked... I wonder whether there are some checks we could add to archetype handling or shape-of constraint handling to fail if code like this is added in the future?

hborla · 2024-01-26T23:10:12Z

This is subsumed by #70227

hborla force-pushed the same-element-requirements branch from e3a2b4c to 882e1e7 Compare August 29, 2023 21:02

slavapestov reviewed Aug 30, 2023

View reviewed changes

lib/AST/RequirementMachine/Rule.cpp Outdated Show resolved Hide resolved

hborla force-pushed the same-element-requirements branch from e06d45e to 68d6a39 Compare August 31, 2023 03:28

slavapestov reviewed Aug 31, 2023

View reviewed changes

lib/AST/RequirementMachine/RewriteSystem.cpp Outdated Show resolved Hide resolved

slavapestov reviewed Aug 31, 2023

View reviewed changes

hborla force-pushed the same-element-requirements branch from 3253208 to 34971b9 Compare September 2, 2023 19:07

hborla marked this pull request as ready for review September 2, 2023 21:18

hborla requested review from xedin and tshortli as code owners September 2, 2023 21:18

hborla added 7 commits September 2, 2023 17:01

[Requirement Machine] Implement same-element requirements.

7537649

[Requirement Machine] Fix the order of requirements that involve pack…

d138b15

… elements.

[Requirement Machine] Use element symbols for concrete same-element r…

20e8e19

…equirements.

[Constraint System] Fix the shape class and context substitiutions for

c4a1135

opened element generic environments containing same-element requirements.

[Requirement Machine] Stricter verification conditions for element sy…

61b2503

…mbols.

[Features] Gate same-element requirements behind an experimental feat…

819556b

…ure flag.

[NFC] Add -dump-input=fail to a FileCheck test.

f381b3f

hborla force-pushed the same-element-requirements branch from 34971b9 to f381b3f Compare September 3, 2023 00:29

slavapestov reviewed Sep 3, 2023

View reviewed changes

xedin reviewed Sep 4, 2023

View reviewed changes

simanerush mentioned this pull request Sep 25, 2023

[SE-0408] Enable Pack Iteration #67594

Merged

13 tasks

simanerush self-assigned this Nov 7, 2023

simanerush mentioned this pull request Dec 5, 2023

[Requirement Machine] Implement same-element requirements. #70227

Merged

3 tasks

hborla closed this Jan 26, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[Requirement Machine] Implement same-element requirements. #67465

[Requirement Machine] Implement same-element requirements. #67465

Uh oh!

hborla commented Jul 22, 2023 •

edited

Loading

Uh oh!

Uh oh!

Uh oh!

slavapestov Aug 31, 2023

Uh oh!

hborla Aug 31, 2023 •

edited

Loading

Uh oh!

slavapestov Aug 31, 2023

Uh oh!

hborla commented Sep 2, 2023

Uh oh!

hborla commented Sep 3, 2023

Uh oh!

slavapestov Sep 3, 2023

Uh oh!

xedin left a comment

Uh oh!

xedin Sep 4, 2023

Uh oh!

xedin Sep 4, 2023

Uh oh!

hborla commented Jan 26, 2024

Uh oh!

Uh oh!

[Requirement Machine] Implement same-element requirements. #67465

[Requirement Machine] Implement same-element requirements. #67465

Uh oh!

Conversation

hborla commented Jul 22, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

slavapestov Aug 31, 2023

Choose a reason for hiding this comment

Uh oh!

hborla Aug 31, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

slavapestov Aug 31, 2023

Choose a reason for hiding this comment

Uh oh!

hborla commented Sep 2, 2023

Uh oh!

hborla commented Sep 3, 2023

Uh oh!

slavapestov Sep 3, 2023

Choose a reason for hiding this comment

Uh oh!

xedin left a comment

Choose a reason for hiding this comment

Uh oh!

xedin Sep 4, 2023

Choose a reason for hiding this comment

Uh oh!

xedin Sep 4, 2023

Choose a reason for hiding this comment

Uh oh!

hborla commented Jan 26, 2024

Uh oh!

Uh oh!

hborla commented Jul 22, 2023 •

edited

Loading

hborla Aug 31, 2023 •

edited

Loading