Fix dotty CI

[hamzaremmal]

Closes #20362

[hamzaremmal]

When I generated the checkfile in the branch of #20325, I get the following feedback

-- Error: tests/neg/i20317.scala:3:10 ----------------------------------------------------------------------------------
3 |def foo[A <: Foo[A]]: Unit = () // error // error
  |          ^
  |          Recursion limit exceeded.
  |          Maybe there is an illegal cyclic reference?
  |          If that's not the case, you could also try to increase the stacksize using the -Xss JVM option.
  |          For the unprocessed stack trace, compile with -Yno-decode-stacktraces.
  |          A recurring operation is (inner to outer):
  |
  |            type parameters of Foo[A]
  |            type parameters of  <: Foo[A]
  |            type parameters of A
  |            type parameters of A
  |            type parameters of Foo[A]
  |            type parameters of  <: Foo[A]
  |            type parameters of A
  |            type parameters of A
  |            type parameters of Foo[A]
  |            type parameters of  <: Foo[A]
  |            ...
  |
  |            type parameters of A
  |            type parameters of Foo[A]
  |            type parameters of  <: Foo[A]
  |            type parameters of A
  |            type parameters of A
  |            type parameters of Foo[A]
  |            type parameters of  <: Foo[A]
  |            type parameters of A
  |            type parameters of Foo[A]
  |            type parameters of  <: Foo[A]
-- Error: tests/neg/i20317.scala:3:8 -----------------------------------------------------------------------------------
3 |def foo[A <: Foo[A]]: Unit = () // error // error
  |        ^
  |        Recursion limit exceeded.
  |        Maybe there is an illegal cyclic reference?
  |        If that's not the case, you could also try to increase the stacksize using the -Xss JVM option.
  |        For the unprocessed stack trace, compile with -Yno-decode-stacktraces.
  |        A recurring operation is (inner to outer):
  |
  |          type parameters of A
  |          type parameters of Foo[A]
  |          type parameters of  <: Foo[A]
  |          type parameters of A
  |          type parameters of A
  |          type parameters of Foo[A]
  |          type parameters of  <: Foo[A]
  |          type parameters of A
  |          type parameters of A
  |          type parameters of Foo[A]
  |          ...
  |
  |          type parameters of A
  |          type parameters of A
  |          type parameters of Foo[A]
  |          type parameters of  <: Foo[A]
  |          type parameters of A
  |          type parameters of A
  |          type parameters of Foo[A]
  |          type parameters of  <: Foo[A]
  |          type parameters of A
  |          type parameters of Foo[A]

@odersky Is this the expected behaviour ? Should this PR be considered as unblocking the CI and not a fix ?

[sjrd]

We've merged so many PRs in the past 2 days that it's not very surprising that we got so-called semantic conflicts. Reporting 1 or 2 "illegal cyclic" errors doesn't make any difference. So if it unblocks the CI, that's good enough.

[odersky]

Yes let's change the check file. It's an evil and completely senseless test problem. We only test that the compiler survives it, with some indication what is wrong.

[hamzaremmal]

@odersky The error reporting is non-deterministic, the CI failed and when I rerun the test locally I get the following checkfile (which is the second error that we thought is ignored):

-- Error: tests/neg/i20317.scala:3:10 ----------------------------------------------------------------------------------
3 |def foo[A <: Foo[A]]: Unit = () // error
  |          ^
  |          Recursion limit exceeded.
  |          Maybe there is an illegal cyclic reference?
  |          If that's not the case, you could also try to increase the stacksize using the -Xss JVM option.
  |          For the unprocessed stack trace, compile with -Xno-decode-stacktraces.
  |          A recurring operation is (inner to outer):
  |
  |            type parameters of  <: Foo[A]
  |            type parameters of A
  |            type parameters of A
  |            type parameters of Foo[A]
  |            type parameters of  <: Foo[A]
  |            type parameters of A
  |            type parameters of A
  |            type parameters of Foo[A]
  |            type parameters of  <: Foo[A]
  |            type parameters of A
  |            ...
  |
  |            type parameters of A
  |            type parameters of Foo[A]
  |            type parameters of  <: Foo[A]
  |            type parameters of A
  |            type parameters of A
  |            type parameters of Foo[A]
  |            type parameters of  <: Foo[A]
  |            type parameters of A
  |            type parameters of Foo[A]
  |            type parameters of  <: Foo[A]

[odersky]

OK, drop the test then. As I said it's a really evil program. I don't care if the behavior is non-deterministic in the number of errors. There's a second test i20317a.scala that behaves better. It's good enough to keep that one.

…

On Wed, May 8, 2024 at 11:01 AM Hamza Remmal ***@***.***> wrote: @odersky <https://github.com/odersky> The error reporting is non-deterministic, the CI failed and when I rerun the test locally I get the following checkfile (which is the second error that we thought is ignored): -- Error: tests/neg/i20317.scala:3:10 ----------------------------------------------------------------------------------3 |def foo[A <: Foo[A]]: Unit = () // error | ^ | Recursion limit exceeded. | Maybe there is an illegal cyclic reference? | If that's not the case, you could also try to increase the stacksize using the -Xss JVM option. | For the unprocessed stack trace, compile with -Xno-decode-stacktraces. | A recurring operation is (inner to outer): | | type parameters of <: Foo[A] | type parameters of A | type parameters of A | type parameters of Foo[A] | type parameters of <: Foo[A] | type parameters of A | type parameters of A | type parameters of Foo[A] | type parameters of <: Foo[A] | type parameters of A | ... | | type parameters of A | type parameters of Foo[A] | type parameters of <: Foo[A] | type parameters of A | type parameters of A | type parameters of Foo[A] | type parameters of <: Foo[A] | type parameters of A | type parameters of Foo[A] | type parameters of <: Foo[A] — Reply to this email directly, view it on GitHub <#20363 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAGCKVWXVFKLDYTRO4U5LLDZBHSYFAVCNFSM6AAAAABHMP3TLGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCMBQGEYDGNJXHA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

-- Martin Odersky Professor, Programming Methods Group (LAMP) Faculty IC, EPFL Station 14, Lausanne, Switzerland