Print wildcard star properly (fix #14495)

[Atry]

No description provided.

[romanowski]

Looks good to me but @odersky or @nicolasstucki should review as well

[smarter]

* instead of _: * is valid syntax in Scala 3.

[odersky]

Yes, it should be "*", not ": *"

[smarter]

A SeqLiteral is a specific kind of tree node, when debugging the compiler output with -Xprint, it's useful to be able to distinguish a SeqLiteral produced by the compiler from a Seq tree written by the user. If this doesn't appear in something that the compiler prints to the user then I don't think it should be changed.

[odersky]

Agree with @smarter. We should at least retain the old output under -Yprint-debug

[Atry]

It appears in compiletime.codeOf Guillaume Martres ***@***.***>于2022年2月19日 周六上午8:51写道：

…

***@***.**** commented on this pull request. ------------------------------ In compiler/src/dotty/tools/dotc/printing/RefinedPrinter.scala <#14496 (comment)>: > @@ -513,7 +513,7 @@ class RefinedPrinter(_ctx: Context) extends PlainPrinter(_ctx) { keywordStr("throw ") ~ toText(expr) } case SeqLiteral(elems, elemtpt) => - "[" ~ toTextGlobal(elems, ",") ~ " : " ~ toText(elemtpt) ~ "]" A SeqLiteral is a specific kind of tree node, when debugging the compiler output with -Xprint, it's useful to be able to distinguish a SeqLiteral produced by the compiler from a Seq tree written by the user. If this doesn't appear in something that the compiler prints to the user then I don't think it should be changed. — Reply to this email directly, view it on GitHub <#14496 (review)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAES3OTQGG3XJ6NTJRDC3DLU37DBFANCNFSM5OQ4CYBA> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[smarter]

OK, I don't know what the intended purpose of codeOf is, but if it's supposed to produce code that compiles, it shouldn't rely on RefinedPrinter only.

[odersky]

Overall, I don't see that this improves things. I believe it would be good to make an effort to get to a printer that only outputs code that can be parsed back.

But there is a natural tension between that and diagnostics that give more information. So rather than fiddling with details I believe we need an overall strategy how we can get to such a printer without duplicating a huge amount of code.

[odersky]

Yes, it should be "*", not ": *"

[odersky]

Agree with @smarter. We should at least retain the old output under -Yprint-debug

[smarter]

We do have showDecompiledTree ( https://github.com/lampepfl/dotty/blob/cf9525c713b99ae4046234f8ff350c97a9d380d0/compiler/src/scala/quoted/runtime/impl/QuotesImpl.scala#L29 ), I don't know if it could be reused for codeOf /cc @nicolasstucki

[nicolasstucki]

showDecompiledTree works for this use case, but the formatting would be a bit different. The code shown would be the fully elaborated code. This implies that all prefixes would be fully expanded.

[ckipp01]

Hey @Atry do you plan on returning back to this?

[ckipp01]

Hey @Atry do you plan on returning back to this?

Alright, since it's been a month and I haven't heard back I'm going to go ahead and close this. Please do return back if you're interested! I'll link #14495 so it can be found later on as well.