Re: [gwt-contrib] Re: GWT 2.8.0 SNAPSHOT compiler crash

2016-04-27 Thread 'Daniel Kurka' via GWT Contributors
With j2cl we do a javac compile up front and thus those basic problems are gone. On Wed, Apr 27, 2016 at 4:22 PM Colin Alworth wrote: > Agreed - validating supersource is tricky. I've heard of some good options > lately that change how the project interacts with them (like

Re: [gwt-contrib] Re: GWT 2.8.0 SNAPSHOT compiler crash

2016-04-27 Thread Colin Alworth
Agreed - validating supersource is tricky. I've heard of some good options lately that change how the project interacts with them (like compiling to .class file and leveraging javac to get the errors), but I don't presently have any best practices to suggest in general, except tests for each

Re: [gwt-contrib] Re: GWT 2.8.0 SNAPSHOT compiler crash

2016-04-27 Thread David
Yes the org.w3c stuff was my own, it was defined in the html5 module you see in the classpath. I hope that this was a border case and not typical what happens if you have super sources with import statements that cannot be resolved ? Because in that case the compiler should give a better error

Re: [gwt-contrib] Re: GWT 2.8.0 SNAPSHOT compiler crash

2016-04-27 Thread Colin Alworth
Thanks. Can you confirm that org.w3c.dom is your own code, and not part of GWT, so it was just out of date? I don't see a jar with that name on your classpath, and neither of the two EntityReference.java files in GWT live in that package or appear have jsinterop annotations. On Wednesday,

Re: [gwt-contrib] Re: GWT 2.8.0 SNAPSHOT compiler crash

2016-04-27 Thread David
Colin, The stacktrace was is my initial post. Anyway: I managed to find the root cause and solution. The super source java file is using JsInterop to map org.w3c.dom API's. It had annotations like this: @JsType(isNative=true, namespace=JsNamespace.GLOBAL). However, in the snapshot this

Re: [gwt-contrib] Re: GWT 2.8.0 SNAPSHOT compiler crash

2016-04-27 Thread Colin Alworth
Can you share the stack trace, the file it is attempting to read, and your classpath? It does sound like it might be a bug, but without the ability to reproduce, it is difficult to say more. One case where it might not be a bug despite its behavior: where it is reusing gwt-unitCache or the like

Re: [gwt-contrib] Re: GWT 2.8.0 SNAPSHOT compiler crash

2016-04-27 Thread David
I managed to get this to run in a Debugger, from within the maven invocation. The exception is thrown on a super source file in one of my gwt-lib maven artifacts. It is trying to parse a .java file as a class file and this throws the reported exception. This sounds like a bug to me!? On Wed, 27

[gwt-contrib] Re: GWT 2.8.0 SNAPSHOT compiler crash

2016-04-27 Thread Jens
Hmm never seen this. Maybe some ASM 3.x has creeped into your classpath? GWT 2.8 requires ASM 5.x. -- J. -- You received this message because you are subscribed to the Google Groups "GWT Contributors" group. To unsubscribe from this group and stop receiving emails from it, send an email to