Improve diagnostic in TDML Runner

Note that this change enforces the syntax of TDML better.
The dfdlInfoset element used to tolerate having spurious text in it,
so long as it also had a root element.
Two TDML files had to be fixed in Daffodil.
If TDML files in schema projects have these spurious characters that
were formerly ignored, this change will break
those tests.

DAFFODIL-533

daffodil-lib/src/main/resources/org/apache/daffodil/xsd/tdml.xsd

@@ -308,7 +308,15 @@
 
   <complexType name="dfdlInfosetType" mixed="true">
     <sequence>
-      <any namespace="##any" processContents="lax" minOccurs="0" maxOccurs="unbounded"/>
+      <!--
+      Can be empty if type is 'file'.
+      Otherwise, can be any single element (maxOccurs unbounded removed),
+      which is the root element of the infoset.
+
+      Mixed="true" because when type='file' the content is just a string
+      that is the file path
+      -->
+      <any namespace="##any" processContents="lax" minOccurs="0" maxOccurs="1"/>
     </sequence>
     <attribute name="type" use="optional" default="infoset">
       <simpleType>

daffodil-tdml-lib/src/main/scala/org/apache/daffodil/tdml/TDMLRunner.scala

@@ -31,6 +31,7 @@ import java.nio.charset.StandardCharsets
 import scala.collection.mutable
 import scala.language.postfixOps
 import scala.util.Try
+import scala.xml.Elem
 import scala.xml.Node
 import scala.xml.NodeSeq
 import scala.xml.NodeSeq.seqToNodeSeq
@@ -2769,41 +2770,52 @@ case class Infoset(i: NodeSeq, parent: TestCase) {
 
 case class DFDLInfoset(di: Node, parent: Infoset) {
 
-  private lazy val infosetNodeSeq = {
-    val testCase: TestCase = parent.parent
-    val loader = testCase.parent.loader
-    val optDataSchema: Option[URI] = {
-      testCase.optSchemaFileURI.orElse(testCase.optEmbeddedSchema.map { _.uriForLoading })
-    }
-    val src =
-      (di \ "@type").toString match {
-        case "infoset" | "" => {
-          val rawElem = di.child.filter {
-            _.isInstanceOf[scala.xml.Elem]
-          }.head
-          UnitTestSchemaSource(rawElem, testCase.tcName)
-        }
-        case "file" => {
-          val path = di.text.trim()
-          val maybeURI = parent.parent.parent.findTDMLResource(path)
-          val uri = maybeURI.getOrElse(
-            throw new FileNotFoundException(
-              "TDMLRunner: infoset file '" + path + "' was not found"
-            )
-          )
-          URISchemaSource(uriToDiagnosticFile(uri), uri)
-        }
-        case value => Assert.abort("Unknown value for type attribute on dfdlInfoset: " + value)
-      }
+  private lazy val testCase: TestCase = parent.parent
+  private lazy val loader = testCase.parent.loader
+  private val ty: String = {
+    val t = (di \ "@type").text.trim
+    if (t.isEmpty) "infoset" else t
+  }
+
+  private val elemOrStr: Either[Elem, String] = {
+    val (elems, others) = di.child.partition(_.isInstanceOf[scala.xml.Elem])
+    (elems, others.text.trim) match {
+      case (Seq(elem: Elem), "") if (ty == "infoset") => Left(elem)
+      case (Seq(), str) if (ty == "file") => Right(str)
+      case _ =>
+        Assert.usageError(
+          """dfdlInfoset element must contain a single root element or a file path (when 'type="file"')."""
+        )
+    }
+  }
+
+  lazy val contents: Elem = {
+    elemOrStr match {
+      case Left(elem) => elem
+      case Right(path) => infosetNodeFromFile(path)
+    }
+  }
+
+  private def infosetNodeFromFile(path: String): Elem = {
+
+    val maybeURI = parent.parent.parent.findTDMLResource(path)
+    val uri = maybeURI.getOrElse(
+      throw new FileNotFoundException(
+        "TDMLRunner: infoset file '" + path + "' was not found"
+      )
+    )
+    val infosetSrc = URISchemaSource(uriToDiagnosticFile(uri), uri)
+
+    val testSuite = testCase.parent
+    val before = testSuite.loadingExceptions.clone()
+
+    val elem = loader.load(infosetSrc, None) // no schema
     //
     // TODO: DAFFODIL-288 validate the infoset also
     // You can pass the optDataSchema, which appears to be the correct thing
     // but in many cases it doesn't seem to be able to resolve things.
     //
-    val testSuite = testCase.parent
-    val before = testSuite.loadingExceptions.clone()
-    // val elem: Node = loader.load(src, optDataSchema)
-    val elem = loader.load(src, None) // no schema
+    // val elem: Node = loader.load(infosetSrc, optDataSchema)
 
     // The expected infoset is loaded using the normalizeCRLFtoLF mode (which
     // is the default), so MS-DOS/Windows CRLFs in expected data XML files will
@@ -2818,17 +2830,9 @@ case class DFDLInfoset(di: Node, parent: Infoset) {
       val newExceptions = (testSuite.loadingExceptions -- before).toSeq
       testCase.toss(TDMLException(newExceptions, None), None)
     }
-    elem
+    elem.asInstanceOf[Elem]
   }
 
-  lazy val contents = {
-    Assert.usage(
-      infosetNodeSeq.size == 1,
-      "dfdlInfoset element must contain a single root element"
-    )
-    val c = infosetNodeSeq.head
-    c
-  }
 }
 
 object DiagnosticType extends Enumeration {

daffodil-tdml-lib/src/test/scala/org/apache/daffodil/tdml/UnitTestTDMLRunner.scala

@@ -21,6 +21,7 @@ import java.io.File
 
 import org.apache.daffodil.lib.Implicits._
 import org.apache.daffodil.lib.Implicits.using
+import org.apache.daffodil.lib.exceptions.UsageException
 import org.apache.daffodil.lib.util._
 import org.apache.daffodil.lib.xml.XMLUtils
 
@@ -336,6 +337,17 @@ class UnitTestTDMLRunner {
     assertEquals(expected, actual.toList)
   }
 
+  @Test def testDFDLInfosetEmptyDiagnosticMsg(): Unit = {
+    val xml = <tdml:dfdlInfoset/>
+    val exc = intercept[UsageException] {
+      val di = new DFDLInfoset(xml, null)
+      di.contents
+    }
+    val m = exc.getMessage
+    assertTrue(m.contains("dfdlInfoset"))
+    assertTrue(m.contains("single root element"))
+  }
+
   @Test def testLSB1(): Unit = {
     val xml = <document bitOrder="LSBFirst">
                 <documentPart type="bits">00000010</documentPart>

daffodil-tdml-processor/src/test/scala/org/apache/daffodil/processor/tdml/TestTDMLRunner.scala

@@ -18,6 +18,7 @@
 package org.apache.daffodil.processor.tdml
 
 import java.io.File
+import java.io.FileNotFoundException
 
 import org.apache.daffodil.lib.Implicits._
 import org.apache.daffodil.lib.Implicits.using
@@ -1061,4 +1062,32 @@ f0 f1 f2 f3 f4 f5 f6 f7 f8 f9 fa fb fc fd fe ff
     assertTrue(exc.getMessage.contains("Either tdml:infoset or tdml:error"))
     assertTrue(exc.getMessage.contains("must be present in the test"))
   }
+
+  @Test def testInfosetFileNotFound() = {
+    val testSuite =
+      <tdml:testSuite suiteName="theSuiteName" xmlns:tdml={tdml} xmlns:dfdl={dfdl}
+                      xmlns:xs={xsd}>
+        <tdml:defineSchema name="mySchema">
+          <xs:include schemaLocation="/org/apache/daffodil/xsd/DFDLGeneralFormat.dfdl.xsd"/>
+          <dfdl:format ref="GeneralFormat"/>
+          <xs:element name="data" type="xs:int" dfdl:lengthKind="delimited"/>
+        </tdml:defineSchema>
+        <tdml:unparserTestCase name="infosetFileNotFound" root="data" model="mySchema">
+          <tdml:infoset>
+            <tdml:dfdlInfoset type="file">/this/does/not/exist.xml</tdml:dfdlInfoset>
+          </tdml:infoset>
+          <tdml:document/>
+        </tdml:unparserTestCase>
+      </tdml:testSuite>
+
+    val runner = new Runner(testSuite)
+    val e = intercept[FileNotFoundException] {
+      runner.runOneTest("infosetFileNotFound")
+    }
+    runner.reset
+    val msg = e.getMessage()
+    assertTrue(msg.contains("not found"))
+    assertTrue(msg.contains("/this/does/not/exist.xml"))
+  }
+
 }

daffodil-test/src/test/resources/org/apache/daffodil/section15/choice_groups/HiddenChoices.tdml

@@ -217,7 +217,7 @@
   <tdml:unparserTestCase name="nestedNoOVCinHiddenContext" root="e7"
     model="ChoicesInHiddenContexts.dfdl.xsd" description="hidden group ref">
     <tdml:infoset>
-      <tdml:dfdlInfoset>)
+      <tdml:dfdlInfoset>
         <ex:e7/>
       </tdml:dfdlInfoset>
     </tdml:infoset>