Implement SIP-67: strictEquality pattern matching (fixes #22732) #23803

mberndt123 · 2025-08-24T21:36:16Z

Hi,

this is my attempt at implementing the SIP-67 specification. Thanks to @noti0na1 and @dwijnand for the hints you gave in #22732. Sorry it took so long, please let me know what you think.

…#22732)

odersky · 2025-08-30T11:05:14Z

compiler/src/dotty/tools/dotc/typer/Typer.scala

-            untpd.Select(untpd.TypedSplice(tree), nme.EQ),
-            untpd.TypedSplice(dummyTreeOfType(pt)))
-        typedExpr(cmp, defn.BooleanType)
+        if ! (tree.tpe <:< pt && (tree.symbol.flags.isAllOf(Flags.EnumValue) || (tree.symbol.flags.isAllOf(Flags.Module | Flags.Case)))) then


Some remarks:

The motivation for the SIP was that it makes strictEquality easier to use, but this also changes the pattern matching behavior of patterns without that import. And it breaks the law that pattern matching values is done with an == check. I think this goes too far, even though this is what's specified in the SIP.

We cannot implement a SIP change without going through an experimenal language import.

I think it would be better to add code to assumedCanEqual instead. Under strictEquality and the experimental import, if the left operand (or either one) is an enum constant, and the other operand is a supertype, assume the two CanEqual.

Hi @odersky,

Thank you for your review.

Regarding 1: 🤔 I'm not sure I understand. The SIP says “The semantics of pattern matching against a constant are otherwise unchanged”, meaning that it will still be equivalent to ==, except it doesn't require CanEqual. If that's not what it does, then it's a bug in my implementation. But I think that's solved by moving it to assumeCanEqual as you suggested in point 3, right?

Regarding 2: ✅ I've added an experimental language import, strictEqualityPatternMatching, to enable this behaviour. I also tried adding a Mode for this so it can be enabled with a command line flag similar to -language:strictEquality, but unless I'm misreading the code, the modes are stored in an Int bitmask and all the bits are already taken 😒

Regarding 3: ✅ I've moved the implementation to assumedCanEqual, which was a bit tricky because in that function the Tree isn't available, but I need it to know if it's a case object or enum case. It's also called from Synthesizer where there is no Tree, so I'm passing it as an Option[Tree] now.

hamzaremmal · 2025-09-15T19:50:24Z

library/src/scala/runtime/stdLibPatches/language.scala

Please add this object in scala.language.experimental too (Not just in the stdLibPatches).

Thanks for pointing this out Hamza, I've now added it.

mberndt123 · 2025-09-23T15:01:47Z

Hey @Gedochao ,

It's been a while since you assigned this to Martin, and I addressed his feedback as well as @hamzaremmal's. But I have the impression that it wasn't noticed yet, perhaps you could assign it back to @odersky?

soronpo · 2025-09-23T15:09:41Z

I will mention the implementation is ready in the upcoming SIP meeting this Friday. It's already on the agenda.

mberndt123 · 2025-09-23T15:11:49Z

Thanks @soronpo!

SethTisue · 2025-09-24T00:12:44Z

tentatively milestoning for 3.8.0 since it should least be evaluated for inclusion

odersky

Otherwise LGTM in the sense that it implements SIP 67 as specced and I believe after the suggested changes are implemented it will fit well with the rest of the compiler.

odersky · 2025-09-24T17:13:14Z

compiler/src/dotty/tools/dotc/typer/Implicits.scala

   *     unless strict equality is set.
   */
-  def assumedCanEqual(ltp: Type, rtp: Type)(using Context) = {
+  def assumedCanEqual(leftTreeOption: Option[Tree], ltp: Type, rtp: Type)(using Context): Boolean = {


We usually use EmptyTree for missing trees instead of an option type. Also, leftTree should be the last argument. Suggestion:

def assumedCanEqual(ltp: Type, rtp: Type, leftTree: Tree = EmptyTree)

Done, thanks 👍🏻

odersky · 2025-09-24T17:14:41Z

compiler/src/dotty/tools/dotc/typer/Implicits.scala

  /** Check that equality tests between types `ltp` and `rtp` make sense */
-  def checkCanEqual(ltp: Type, rtp: Type, span: Span)(using Context): Unit =
-    if (!ctx.isAfterTyper && !assumedCanEqual(ltp, rtp)) {
+  def checkCanEqual(left: Tree, rtp: Type, span: Span)(using Context): Unit =


ltp is not part of the API anymore, so the doc comment needs to be updated.

Thanks, I've updated the comment

odersky · 2025-09-24T17:23:26Z

compiler/src/dotty/tools/dotc/typer/Implicits.scala

+    || locally:
+      if strictEquality then
+        strictEqualityPatternMatching && leftTreeOption.exists: leftTree =>
+          ltp <:< lift(rtp) && (leftTree.symbol.flags.isAllOf(Flags.EnumValue) || leftTree.symbol.flags.isAllOf(Flags.Module | Flags.Case))


.flags is redundant, should be omitted. Suggestion:

if strictEqualityPatternMatching && (leftTree.symbol.isAllOf(Flags.EnumValue) || leftTree.symbol.isAllOf(Flags.Module | Flags.Case)) then ltp <:< lift(rtp) else false

More pedestrian but much clearer and more efficient.

Thanks, I've removed the .flags. Re replacing && with if: that would lead to nested if expressions, and I think a simple && is easier to read than that. I'd also like to understand why if would be more efficient given that && implements short-circuiting.

I'd prefer leaving it like this, but will change it if you insist.

tgodzik · 2025-09-29T09:24:02Z

@odersky is it ok to merge now?

odersky · 2025-10-02T17:38:34Z

Yes, let's merge.

odersky · 2025-10-02T17:40:30Z

@hamzaremmal You also need to approve before we can merge.

mberndt123 force-pushed the sip-67-strict-equality-improvements branch from 3bf5df8 to e3d22ab Compare August 26, 2025 21:14

mberndt123 added 2 commits August 26, 2025 23:14

Implement SIP-67: strictEquality pattern matching (Github issue scala…

5778bea

…#22732)

use braceless syntax in SIP67Tests

0a565d0

mberndt123 force-pushed the sip-67-strict-equality-improvements branch from e3d22ab to 0a565d0 Compare August 26, 2025 21:14

Gedochao requested review from noti0na1 and odersky August 27, 2025 11:19

Gedochao assigned odersky and noti0na1 Aug 27, 2025

odersky requested changes Aug 30, 2025

View reviewed changes

odersky assigned mberndt123 and unassigned odersky Aug 30, 2025

mberndt123 requested a review from a team as a code owner August 31, 2025 20:54

mberndt123 force-pushed the sip-67-strict-equality-improvements branch 3 times, most recently from 2bb5197 to e9fb60b Compare August 31, 2025 21:17

mberndt123 added 2 commits August 31, 2025 23:45

change SIP-67 implementation based on @odersky's review

d433ebe

hide SIP-67 behaviour behind experimental feature flag

4f7a9a3

mberndt123 force-pushed the sip-67-strict-equality-improvements branch from e9fb60b to 4f7a9a3 Compare August 31, 2025 21:46

mberndt123 requested a review from odersky August 31, 2025 22:06

mberndt123 added 3 commits September 1, 2025 01:00

remove excess parens

56f11c2

minor refactoring of SIP67Tests

6570fb2

whitespace fix

51b34cb

hamzaremmal requested changes Sep 15, 2025

View reviewed changes

mberndt123 requested a review from hamzaremmal September 15, 2025 21:26

add strictEqualityPatternMatching in scala.language.experimental

79fc91d

mberndt123 force-pushed the sip-67-strict-equality-improvements branch from a82ddbd to 79fc91d Compare September 15, 2025 21:28

Merge branch 'main' into sip-67-strict-equality-improvements

45e188c

SethTisue added this to the 3.8.0 milestone Sep 24, 2025

odersky requested changes Sep 24, 2025

View reviewed changes

address review comments

f23deba

mberndt123 requested a review from odersky September 26, 2025 11:14

mberndt123 added 3 commits September 26, 2025 13:15

Merge branch 'main' into sip-67-strict-equality-improvements

341e927

comments

684663a

Merge branch 'main' into sip-67-strict-equality-improvements

5f62655

odersky approved these changes Oct 2, 2025

View reviewed changes

hamzaremmal approved these changes Oct 2, 2025

View reviewed changes

odersky merged commit 172687a into scala:main Oct 2, 2025
49 checks passed

Implement SIP-67: strictEquality pattern matching (fixes #22732) #23803

Implement SIP-67: strictEquality pattern matching (fixes #22732) #23803

Uh oh!

Conversation

mberndt123 commented Aug 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

odersky Aug 30, 2025 • 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!

mberndt123 Aug 31, 2025 • 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!

hamzaremmal Sep 15, 2025

Choose a reason for hiding this comment

Uh oh!

mberndt123 Sep 15, 2025 • 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!

mberndt123 commented Sep 23, 2025

Uh oh!

soronpo commented Sep 23, 2025

Uh oh!

mberndt123 commented Sep 23, 2025

Uh oh!

SethTisue commented Sep 24, 2025

Uh oh!

odersky left a comment

Choose a reason for hiding this comment

Uh oh!

odersky Sep 24, 2025

Choose a reason for hiding this comment

Uh oh!

mberndt123 Sep 25, 2025

Choose a reason for hiding this comment

Uh oh!

odersky Sep 24, 2025

Choose a reason for hiding this comment

Uh oh!

mberndt123 Sep 25, 2025

Choose a reason for hiding this comment

Uh oh!

odersky Sep 24, 2025

Choose a reason for hiding this comment

Uh oh!

mberndt123 Sep 25, 2025 • 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!

tgodzik commented Sep 29, 2025

Uh oh!

odersky commented Oct 2, 2025

Uh oh!

odersky commented Oct 2, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

mberndt123 commented Aug 24, 2025 •

edited

Loading

odersky Aug 30, 2025 •

edited

Loading

mberndt123 Aug 31, 2025 •

edited

Loading

mberndt123 Sep 15, 2025 •

edited

Loading

mberndt123 Sep 25, 2025 •

edited

Loading