The really strange thing is that the virtual machine that launches the
dos command does indeed
spawn correctly as I can see its process complete. I've even debugged
it right to the very end.
So it seems that the grandparent process responsible for running the
ant java task has some strange
Didn't make any difference. I did however get the stack trace from the
ant java task vm and the only non daemon thread
dump looks like this:
Thread [main] (Suspended)
ProcessImpl.waitFor() line: not available [native method]
Execute.waitFor(Process) line: 551
The ant-java task needs to be inserted in order to capture the return
code of whatever "body" it executes
and report back to the user. That's why spawn is set to false. Also
fork is set to true so that I can reconfigure
all sorts of options such as the maximum amount of memory for the body.
U
in your programmatic call what happens if you put in a
execTask.setFailonerror(true);
execTask.setFailIfExecutionFails(true);
http://api.dpml.net/ant/1.7.0/org/apache/tools/ant/taskdefs/ExecTask.html#setFailIfExecutionFails(boolean)
before the execute()
public static void main(String[] args) t
Can you insert the Ant project that calls both the Java application and the
batch job?
If so, then you can use the parallel task, see
http://ant.apache.org/manual/CoreTasks/parallel.html
On Mon, Apr 27, 2009 at 12:42 PM, Ashley Williams wrote:
> Ok here's a little more context.
>
> It seems tha
Ok here's a little more context.
It seems that the exec/spawn=true command works as expected when
called on its own directly from java.exe
but not from an intermediate ant java task, so:
1. GOOD... When I launch the cmd.exe from a programmatic ant exec
task, the exec task and java.exe both
