FULL PRODUCT VERSION :
java version "10-ea"
Java(TM) SE Runtime Environment (build 10-ea+33)
Java HotSpot(TM) 64-Bit Server VM (build 10-ea+33, mixed mode)


A DESCRIPTION OF THE PROBLEM :
Compiler in JDK 10-ea+33 misses to include entry in LineNumberTable for goto instruction of foreach loop. Due to this in case if body of a loop is not executed, debugger will misleadingly show that step after line of foreach loop will be a last line of the body.

Background information: I'm one of developers of the JaCoCo code coverage tool. This tool relies on line number debug information within class files to provide code coverage highlighting. We have a regression test suite which uncovered the defect reported here.

And due to this issue in the same case code coverage tool such as JaCoCo will highlight misleadingly last line of a body of foreach loop as partially executed.

After bisection of http://hg.openjdk.java.net/jdk/jdk/ seems that this regression was introduced by changes for JDK-8175883 :

$ hg bisect --good jdk-10+30
$ hg bisect --bad jdk-10+33
Testing changeset 47839:314ac2e2db63 (143 changesets remaining, ~7 tests)
1133 files updated, 0 files merged, 320 files removed, 0 files unresolved

$ hg bisect --good
Testing changeset 47845:2ac93efc62ed (71 changesets remaining, ~6 tests)
685 files updated, 0 files merged, 33 files removed, 0 files unresolved

$ hg bisect --good 
Testing changeset 47862:ebe854b910ff (35 changesets remaining, ~5 tests)
93 files updated, 0 files merged, 7 files removed, 0 files unresolved

$ hg bisect --good                             
Testing changeset 47871:5ab3961d20dd (18 changesets remaining, ~4 tests)
51 files updated, 0 files merged, 1 files removed, 0 files unresolved

$ hg bisect --bad                              
Testing changeset 47866:39db80b32b69 (9 changesets remaining, ~3 tests)
29 files updated, 0 files merged, 4 files removed, 0 files unresolved

$ hg bisect --bad                              
Testing changeset 47864:e29ae57c6421 (4 changesets remaining, ~2 tests)
17 files updated, 0 files merged, 0 files removed, 0 files unresolved

$ hg bisect --bad                              
Testing changeset 47863:abe64cd683a4 (2 changesets remaining, ~1 tests)
2 files updated, 0 files merged, 0 files removed, 0 files unresolved

$ hg bisect --good
The first bad revision is:
changeset:   47864:e29ae57c6421
user:        vromero
date:        Mon Nov 20 17:07:21 2017 -0500
summary:     8175883: bytecode generated for the enhanced for loop may block memory garbage collecting


STEPS TO FOLLOW TO REPRODUCE THE PROBLEM :
Compile the provided source file.

EXPECTED VERSUS ACTUAL BEHAVIOR :
EXPECTED -
"line 5" in LineNumberTable as this is the case with JDK 1.8.0_152-b16 , 9.0.1+11 and 10-ea+30 :

  static void example();
    descriptor: ()V
    flags: (0x0008) ACC_STATIC
    Code:
      stack=1, locals=2, args_size=0
         0: invokestatic  #2                  // Method java/util/Collections.emptyList:()Ljava/util/List;
         3: invokeinterface #3,  1            // InterfaceMethod java/util/List.iterator:()Ljava/util/Iterator;
         8: astore_0
         9: aload_0
        10: invokeinterface #4,  1            // InterfaceMethod java/util/Iterator.hasNext:()Z
        15: ifeq          32
        18: aload_0
        19: invokeinterface #5,  1            // InterfaceMethod java/util/Iterator.next:()Ljava/lang/Object;
        24: astore_1
        25: aload_1
        26: invokestatic  #6                  // Method nop:(Ljava/lang/Object;)V
        29: goto          9
        32: return
      LineNumberTable:
        line 3: 0
        line 4: 25
        line 5: 29
        line 6: 32
      StackMapTable: number_of_entries = 2
        frame_type = 252 /* append */
          offset_delta = 9
          locals = [ class java/util/Iterator ]
        frame_type = 250 /* chop */
          offset_delta = 22

ACTUAL -
No "line 5" in LineNumberTable with JDK 10-ea+33 :

  static void example();
    descriptor: ()V
    flags: (0x0008) ACC_STATIC
    Code:
      stack=1, locals=2, args_size=0
         0: invokestatic  #2                  // Method java/util/Collections.emptyList:()Ljava/util/List;
         3: invokeinterface #3,  1            // InterfaceMethod java/util/List.iterator:()Ljava/util/Iterator;
         8: astore_0
         9: aload_0
        10: invokeinterface #4,  1            // InterfaceMethod java/util/Iterator.hasNext:()Z
        15: ifeq          32
        18: aload_0
        19: invokeinterface #5,  1            // InterfaceMethod java/util/Iterator.next:()Ljava/lang/Object;
        24: astore_1
        25: aload_1
        26: invokestatic  #6                  // Method nop:(Ljava/lang/Object;)V
        29: goto          9
        32: aconst_null
        33: astore_0
        34: aconst_null
        35: astore_1
        36: return
      LineNumberTable:
        line 3: 0
        line 4: 25
        line 6: 36
      StackMapTable: number_of_entries = 2
        frame_type = 252 /* append */
          offset_delta = 9
          locals = [ class java/util/Iterator ]
        frame_type = 22 /* same */


REPRODUCIBILITY :
This bug can be reproduced always.

---------- BEGIN SOURCE ----------
class Example {
  static void example() {
    for (Object o : java.util.Collections.emptyList()) {
      nop(o);
    }
  }

  static void nop(Object o) {
  }
}

---------- END SOURCE ----------