Scaladoc member permalinks now get us to destination, not to neighbors

[zerosum]

Fixes scala/bug#11805 .

This problem seems to be just because the fixed header which is identified with search had be hiding other elements.

[zerosum]

I've signed the CLA.
Should I do any other actions?

[som-snytt]

/sync

[eed3si9n]

@zerosum Could you post screenshots of before and after to demonstrate your fix similar to scala/bug#11805 (comment) please?

[zerosum]

Before

1. Open https://www.scala-lang.org/api/current/scala/PartialFunction.html (Scala 2.13.3) using Chrome 86 or Safari 14.0. Then do below steps:

• Scroll down to def andThen[C](k: PartialFunction[B, C]): PartialFunction[A, C], right-click on the link icon, and open a new tab.
• Repeat right-click, new-tab for applyOrElseA1<:A,B1>:B:B1.
• Repeat right-click, new-tab for unlift:PartialFunction[A,B]

andThen[C](k: PartialFunction[B, C]): PartialFunction[A, C]

Chrome, new-tab

Safari, new-tab

applyOrElse[A1<:A,B1>:B](x:A1,default:A1=>B1):B1

Chrome, new-tab

Safari, new-tab

unlift:PartialFunction[A,B]

Chrome, new-tab

Safari, new-tab

After

steps:

1. checkout 6112b18
2. sbt dist/mkPack to build scaladoc
3. generate docs by build/pack/bin/scaladoc -d build/scaladoc/ ./src/library/scala/*.scala
4. Open generated docs using Chrome 86 or Safari 14.0. Then do steps below:

• Scroll down to def andThen[C](k: PartialFunction[B, C]): PartialFunction[A, C], right-click on the link icon, and open a new tab.
• Repeat right-click, new-tab for applyOrElseA1<:A,B1>:B:B1.

// member unlift haven't appears in docs by above steps.

andThen[C](k: PartialFunction[B, C]): PartialFunction[A, C]

Chrome, new-tab

Safari, new-tab

applyOrElse[A1<:A,B1>:B](x:A1,default:A1=>B1):B1

Chrome, new-tab

Safari, new-tab

note

It seems that #9120 solved the problem below:

I'm guessing that this has something to do with < and >?

(via. scala/bug#11805 (comment))

[eed3si9n]

LGTM

[lrytz]

👍 thank you!

[lrytz]

should we make this path a little bit less precise? it seems that it could quite easily break when refactoring anything on that path.

it seems to me, a plain #anchorToMember would be fine.

[zerosum]

you are right. i didn't care about refactoring too much.

i'll modify it. however, i think it would be .anchorToMember, not #anchorToMember.

[lrytz]

Thanks!

[raboof]

Love it! Thanks!