ZoneOffsetTransition state changing during image generation

[galderz]

I've been experimenting with a branch that adds 3rd party library sources into the source cache (#10094), and that has triggered an issue similar to #9750.

Caused by: com.oracle.svm.core.util.UserError$UserException: Static field or an object referenced from a static field changed during native image generation?
  object:2014=[Ljava.time.zone.ZoneOffsetTransition;@51cbb5d  of class: java.util.concurrent.ConcurrentHashMap$Node
  reachable through:
    object: [Ljava.util.concurrent.ConcurrentHashMap$Node;@78ccb00c  of class: java.util.concurrent.ConcurrentHashMap$Node[]
    object: {2017=[Ljava.time.zone.ZoneOffsetTransition;@37d6862f, 2018=[Ljava.time.zone.ZoneOffsetTransition;@66d8e07b, 2019=[Ljava.time.zone.ZoneOffsetTransition;@17730b7e, 2020=[Ljava.time.zone.ZoneOffsetTransition;@58db713f, 2014=[Ljava.time.zone.ZoneOffsetTransition;@51cbb5d}  of class: java.util.concurrent.ConcurrentHashMap
    object: ZoneRules[currentStandardOffset=Z]  of class: java.time.zone.ZoneRules
    object: Europe/London  of class: java.time.ZoneRegion
    object: sun.util.calendar.ZoneInfo[id="Europe/London",offset=0,dstSavings=3600000,useDaylight=true,transitions=242,lastRule=java.util.SimpleTimeZone[id=Europe/London,offset=0,dstSavings=3600000,useDaylight=true,startYear=0,startMode=2,startMonth=2,startDay=-1,startDayOfWeek=1,startTime=3600000,startTimeMode=2,endMode=2,endMonth=9,endDay=-1,endDayOfWeek=1,endTime=3600000,endTimeMode=2]]  of class: sun.util.calendar.ZoneInfo
    object: 0000-12-30T16:47:04.192Z  of class: sun.util.calendar.ImmutableGregorianDate
    object: [CommonEra (C.E.) since 0000-12-30T16:47:04.192 local time]  of class: sun.util.calendar.Era
    object: [Lsun.util.calendar.Era;@225671db  of class: sun.util.calendar.Era[]
    root: sun.util.calendar.JulianCalendar$Date.setNormalizedYear(int)

	at com.oracle.svm.core.util.UserError.abort(UserError.java:68)
	at com.oracle.svm.hosted.image.NativeImageHeapWriter.verifyTargetDidNotChange(NativeImageHeapWriter.java:130)
	at com.oracle.svm.hosted.image.NativeImageHeapWriter.writeReference(NativeImageHeapWriter.java:158)
	at com.oracle.svm.hosted.image.NativeImageHeapWriter.write(NativeImageHeapWriter.java:147)
	at com.oracle.svm.hosted.image.NativeImageHeapWriter.writeConstant(NativeImageHeapWriter.java:184)
	at com.oracle.svm.hosted.image.NativeImageHeapWriter.writeObject(NativeImageHeapWriter.java:377)
	at com.oracle.svm.hosted.image.NativeImageHeapWriter.writeHeap(NativeImageHeapWriter.java:89)
	at com.oracle.svm.hosted.image.NativeBootImage.build(NativeBootImage.java:483)
	at com.oracle.svm.hosted.NativeImageGenerator.doRun(NativeImageGenerator.java:636

Longer output.

[galderz]

Indeed this is strictly related to the debug info integration. The issue happens because of a side effect of the code that inspects the files to add to source cache when inspecting a source root that is a -source.jar. When it does that, SourceCache invokes Files.isRegularFile which indirectly fills up the Java Time ZoneRules.lastRulesCache. When the image is written, the code discovers this change and complains. Not sure about the best fix yet.

[galderz]

FYI, the stacktrace to modifying lastRulesCache can be found here.

[galderz]

This is a GraalVM issue, not a Quarkus one. Closing this, I'll open an issue in GraalVM shortly.