Re: [abc-dev] Pass scheduling

From: Wouter De Borger <>
Date: Fri, 15 Aug 2008 12:41:48 +0200

I patched up the errors,

most where due to the error propagation going wrong

one remains unsolved, it compiles but it doesn't run,

I have some remarks on the whole error reporting thing:

-Errorqueue and Exceptions are interacting in strange ways, I cleaned it up
a little, but it still is odd
-There is absolutely no logic in the class hierarchy of the errors, at
several places I just have to use many catch blocks to prevent certain
exception from being wrapped,
  It would be nice if there was a hierarchy of exceptions so that there
would be only one pass-through catch block
-the actual configuration of the errorqueue is only available after the
polyglot config is parsed. It would be better to have a correct errorqueue
upon entry of runSequence
-I have deduced that an ErrorLimitError should be converted to a
CompilerFailedException, is this correct?
-The runsequence method is spilt into two try-catch blocks, that both wrap
errors, does this have a reason? Is it not sufficient to catch/ wrap all
exceptions in


Running test 792: new/arrays via Class.forName()
Commandline: abc -d new -warn-unused-advice:off new/
not found, skipping...
execution(void ClassForName.main(ClassNotFoundException))
not found, skipping...
InvocationTargetException while trying to run compiled class:
java.lang.ClassNotFoundException: [Ljava.lang.String;
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at java.lang.reflect.Method.invoke(
    at abc.testing.TestCase.runTest(
    at abc.testing.Main.doCase(
    at abc.testing.Main.main(
Caused by: java.lang.ClassNotFoundException: [Ljava.lang.String;
    at java.lang.ClassLoader.findClass(
    at abc.testing.TestCase$FileClassLoader.findClass(
    at java.lang.ClassLoader.loadClass(
    at java.lang.ClassLoader.loadClass(
    at ClassForName.main(
    ... 7 more
FAIL: Test 792: "new/arrays via Class.forName()" failed in 4137ms, memory
usage: 8405160.
Received on Fri Aug 15 2008 - 11:41:52 BST

This archive was generated by hypermail 2.2.0 : Mon Aug 18 2008 - 14:30:11 BST