[Javapathfinder-devel] SF.net SVN: javapathfinder: [523] trunk

[<pcm...@us...>]

Revision: 523
          http://javapathfinder.svn.sourceforge.net/javapathfinder/?rev=523&view=rev
Author:   pcmehlitz
Date:     2007-08-23 00:25:40 -0700 (Thu, 23 Aug 2007)

Log Message:
-----------
* INVOKECG isn't really an InvokeInstruction, it creates one. Subtle difference with
far reaching consequences for listeners on INVOKEs

* The generic Instruction.isCompleted(ti) didn't work for InvokeInstructions, since the
INVOKE is always still on the caller (i.e. lower) stack frame. Now the InvokeInstruction
overrides it, checking if the topframe pc is the first insn in the called method

* TSVM now falls back to the INVOKECG if it finds a test candidate. No need to
look for a different path for the same candidate, since the created test would be the same

* TSVM ignores all java.* exceptions, since those are most likely caused by placeholder
("symbolic") values

Modified Paths:
--------------
    trunk/extensions/symts/src/gov/nasa/jpf/ts/TSVM.java
    trunk/extensions/symts/test/gov/nasa/jpf/ts/TestUnprotectedRW.java
    trunk/src/gov/nasa/jpf/jvm/bytecode/INVOKECG.java
    trunk/src/gov/nasa/jpf/jvm/bytecode/InvokeInstruction.java
    trunk/src/gov/nasa/jpf/tools/CallMonitor.java


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.