[jira] [Updated] (EXEC-71) Changes for EXEC-34 (Race condition prevent watchdog working using ExecuteStreamHandler) hang endlessly if the command cannot be found

["Siegfried Goeschl (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/EXEC-71?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Siegfried Goeschl updated EXEC-71:
----------------------------------
            Flags: Important
           Labels: patch  (was: )
    Fix Version/s: 1.4

> Changes for EXEC-34 (Race condition prevent watchdog working using ExecuteStreamHandler) hang endlessly if the command cannot be found
> --------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: EXEC-71
>                 URL: https://issues.apache.org/jira/browse/EXEC-71
>             Project: Commons Exec
>          Issue Type: Bug
>    Affects Versions: 1.2
>            Reporter: Dominik Stadler
>            Assignee: Siegfried Goeschl
>              Labels: patch
>             Fix For: 1.4
>
>
> The following test hangs endlessly for me because the ExecuteWatchdog will never get "processStarted = true" because starting an non-existing process will never lead to the started flag.
> {code}
> @Test
> 	public void testCommonsExecHang() throws Exception {
>         CommandLine command = new CommandLine("someunexistingcommand");
>         DefaultExecutor executor = new DefaultExecutor();
>         ExecuteWatchdog watchdog = new ExecuteWatchdog(ExecuteWatchdog.INFINITE_TIMEOUT);
>         executor.setWatchdog(watchdog);
>         DefaultExecuteResultHandler resultHandler = new DefaultExecuteResultHandler() {
> 			@Override
> 			public void onProcessFailed(ExecuteException e) {
> 		        System.out.println("Process did not stop gracefully, had exception '" + e.getMessage() + "' while executing process");
> 				super.onProcessFailed(e);
> 			}
>         };
>         executor.execute(command, null, resultHandler);
>         // this hangs!!
>         watchdog.isWatching();
>         watchdog.destroyProcess();
> 	}
> {code}
> not calling isWatching(), but only destroyProcess() hangs as well!
> Options that I see: timeout the loop or handle the case of failure to start the Watchdog in DefaultExecutor.executeInternal() so that the Watchdog does not hang in this case



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)