scala/scala-xml

ConstructingParser throws NoSuchMethodError in Scala 3

cayhorstmann opened this issue · 6 comments

cayhorstmann commented

Save https://horstmann.com/unblog/2020-11-27/index.html to a local file and run

import java.io.File
import scala.xml.parsing.ConstructingParser
val filename="..."
val file = new File(filename)
val parser = ConstructingParser.fromFile(file, preserveWS = true)
val doc = parser.document().docElem

In Scala 2, this works perfectly. In Scala 3.0.0, an exception is thrown:

java.lang.NoSuchMethodError: 'scala.xml.parsing.MarkupParser scala.xml.parsing.MarkupParser$WithLookAhead.scala$xml$parsing$MarkupParser$WithLookAhead$$$outer()'
  at scala.xml.parsing.MarkupParser.lookahead(MarkupParser.scala:77)
  at scala.xml.parsing.MarkupParser.lookahead$(MarkupParser.scala:33)
  at scala.xml.parsing.ConstructingParser.lookahead(ConstructingParser.scala:52)
  at scala.xml.parsing.MarkupParserCommon.peek(MarkupParserCommon.scala:267)
  at scala.xml.parsing.MarkupParserCommon.xTakeUntil(MarkupParserCommon.scala:249)
  at scala.xml.parsing.MarkupParserCommon.xTakeUntil$(MarkupParserCommon.scala:25)
  at scala.xml.parsing.ConstructingParser.xTakeUntil(ConstructingParser.scala:52)
  at scala.xml.parsing.MarkupParser.xCharData(MarkupParser.scala:382)
  at scala.xml.parsing.MarkupParser.xCharData$(MarkupParser.scala:33)
  at scala.xml.parsing.ConstructingParser.xCharData(ConstructingParser.scala:52)
  at scala.xml.parsing.MarkupParser.content1(MarkupParser.scala:427)
  at scala.xml.parsing.MarkupParser.content1$(MarkupParser.scala:33)
  at scala.xml.parsing.ConstructingParser.content1(ConstructingParser.scala:52)
  at scala.xml.parsing.MarkupParser.content(MarkupParser.scala:462)
  at scala.xml.parsing.MarkupParser.content$(MarkupParser.scala:33)
  at scala.xml.parsing.ConstructingParser.content(ConstructingParser.scala:52)
  at scala.xml.parsing.MarkupParser.element1(MarkupParser.scala:591)
  at scala.xml.parsing.MarkupParser.element1$(MarkupParser.scala:33)
  at scala.xml.parsing.ConstructingParser.element1(ConstructingParser.scala:52)
  at scala.xml.parsing.MarkupParser.content1(MarkupParser.scala:436)
  at scala.xml.parsing.MarkupParser.content1$(MarkupParser.scala:33)
  at scala.xml.parsing.ConstructingParser.content1(ConstructingParser.scala:52)
  at scala.xml.parsing.MarkupParser.content(MarkupParser.scala:462)
  at scala.xml.parsing.MarkupParser.content$(MarkupParser.scala:33)
  at scala.xml.parsing.ConstructingParser.content(ConstructingParser.scala:52)
  at scala.xml.parsing.MarkupParser.element1(MarkupParser.scala:591)
  at scala.xml.parsing.MarkupParser.element1$(MarkupParser.scala:33)
  at scala.xml.parsing.ConstructingParser.element1(ConstructingParser.scala:52)
  at scala.xml.parsing.MarkupParser.content1(MarkupParser.scala:436)
  at scala.xml.parsing.MarkupParser.content1$(MarkupParser.scala:33)
  at scala.xml.parsing.ConstructingParser.content1(ConstructingParser.scala:52)
  at scala.xml.parsing.MarkupParser.content(MarkupParser.scala:462)
  at scala.xml.parsing.MarkupParser.content$(MarkupParser.scala:33)
  at scala.xml.parsing.ConstructingParser.content(ConstructingParser.scala:52)
  at scala.xml.parsing.MarkupParser.element1(MarkupParser.scala:591)
  at scala.xml.parsing.MarkupParser.element1$(MarkupParser.scala:33)
  at scala.xml.parsing.ConstructingParser.element1(ConstructingParser.scala:52)
  at scala.xml.parsing.MarkupParser.content1(MarkupParser.scala:436)
  at scala.xml.parsing.MarkupParser.content1$(MarkupParser.scala:33)
  at scala.xml.parsing.ConstructingParser.content1(ConstructingParser.scala:52)
  at scala.xml.parsing.MarkupParser.content(MarkupParser.scala:462)
  at scala.xml.parsing.MarkupParser.content$(MarkupParser.scala:33)
  at scala.xml.parsing.ConstructingParser.content(ConstructingParser.scala:52)
  at scala.xml.parsing.MarkupParser.document(MarkupParser.scala:251)
  at scala.xml.parsing.MarkupParser.document$(MarkupParser.scala:33)
  at scala.xml.parsing.ConstructingParser.document(ConstructingParser.scala:52)

The offending code is

def lookahead(): BufferedIterator[Char] = curInput match {
    case curInputWLA: WithLookAhead =>
      curInputWLA.lookahead()

The library dependency in build.sbt:

libraryDependencies ++= Seq(
  "org.scala-lang.modules" %% "scala-xml" % "2.0.0",
)
SethTisue commented

I'm able to reproduce the problem locally in the Scala 3 REPL.

Once we dig into this it seems like it has to ultimately be a Scala 3 bug rather than something we did wrong in this repo. Regardless, we need this scala-xml ticket anyway because we'll need to publish a new version of the library once there's a compiler fix.

joroKr21 commented

Next step would be to getDeclaredMethods of the WithLookAhead class and see what is the name of the outer accessor (if it exists)

SethTisue commented

minimized (not sure the string is absolutely minimal, but it doesn't matter, we just want something short that triggers it)

  @Test
  def issue541: Unit = {
    val xml =
      """|<script>// <![CDATA[
         |[]; // ]]>
         |</script>""".stripMargin
    val parser = ConstructingParser.fromSource(Source.fromString(xml), preserveWS = true)
    parser.document().docElem  // shouldn't crash
  }
SethTisue commented

I've reported the Scala 3 compiler bug responsible over at scala/scala3#13096

Here in this repo, I will submit a pull request with a workaround.

SethTisue commented

fixed in 2.0.1 which is now published