Context receivers give ParseError

[bddckr]

Using com.diffplug.spotless:spotless-plugin-gradle:6.18.0 configured to use ktfmt().kotlinlangStyle() fails to parse usage of context receivers.

Minimum example:

class A {
    context(String)
    fun test() {}
}

leads to

Step 'ktfmt' found problem in 'A.kt':
2:5: error: Expecting member declaration
com.facebook.ktfmt.format.ParseError: 2:5: error: Expecting member declaration
	at com.facebook.ktfmt.format.Parser.throwParseError(Parser.kt:71)
	at com.facebook.ktfmt.format.Parser.parse(Parser.kt:65)
	at com.facebook.ktfmt.format.Formatter.sortedAndDistinctImports(Formatter.kt:143)
	at com.facebook.ktfmt.format.Formatter.format(Formatter.kt:90)
	at com.diffplug.spotless.glue.ktfmt.KtfmtFormatterFunc.apply(KtfmtFormatterFunc.java:54)
	at com.diffplug.spotless.FormatterFunc.apply(FormatterFunc.java:32)
	at com.diffplug.spotless.FormatterStepImpl$Standard.format(FormatterStepImpl.java:82)
	at com.diffplug.spotless.FormatterStep$Strict.format(FormatterStep.java:88)
	at com.diffplug.spotless.Formatter.compute(Formatter.java:246)
	at com.diffplug.spotless.PaddedCell.calculateDirtyState(PaddedCell.java:203)
	at com.diffplug.gradle.spotless.IdeHook.performHook(IdeHook.java:58)
	at com.diffplug.gradle.spotless.SpotlessExtensionImpl.lambda$createFormatTasks$9(SpotlessExtensionImpl.java:86)

---

According to the context receivers proposal here, the earliest Kotlin release that supports them is 1.6.20-M1. However, ktfmt seems to be using 1.6.10 as seen here. Am I correct to assume that that's the issue?

---

#314 seems related, but according to the comments in that issue, ktfmt at one point worked when using context receivers. This issue here instead explains that ktfmt is basically crashing due to an issue parsing the source code.

[bddckr]

Adding this to TokenizerTest shows the issue:

@Test
fun `Context receivers are parsed correctly`() {
  val code = """
    |class A {
    |  context(String)
    |  fun test() {}
    |}
    |""".trimMargin().trimMargin()

  val file = Parser.parse(code)
  val tokenizer = Tokenizer(code, file)
  file.accept(tokenizer)

  assertThat(tokenizer.toks.map { it.originalText })
          .containsExactly("class", " ", "A", " ", "{", "\n", "  ", "context", "(", "String", ")", "\n", "  ", "fun", " ", "test", "(", ")", " ", "{", "}", "\n", "}")
          .inOrder()
  assertThat(tokenizer.toks.map { it.index })
          .containsExactly(0, -1, 1, -1, 2, -1, -1, 3, 4, 5, 6, -1, -1, 7, -1, 8, 9, 10, -1, 11, 12, -1, 13)
          .inOrder()
}

Changing pom.xml to use <kotlin.version>1.6.20-M1</kotlin.version> makes that test pass.

However, this is where my knowledge ends 😅 Adding this to FormatterTest shows that further changes are required to support context receivers, just like #314 explains.

@Test
fun `context receivers`() =
  assertFormatted(
      """
    |class A {
    |  context(String)
    |  fun test() {}
    |}
    |"""
          .trimMargin())

[davidtorosyan]

@bddckr I don't think #314 indicates that ktfmt used to support context receivers.

Rather, any situations in which they work right now are incidental. In that issue, they're parsed but not formatted correctly.

In your example, they're not parsed at all.

---

Normally I'd close this as a duplicate, but the fact that this causes a crash (instead of just bad formatting) means we should consider a mitigation at least (if a full fix is hard).

@strulovich has there been any investigation into this?