Session.lock(object, LockMode.READ) prints annoying StackTra

bertrand · **Posted:** Mon Jan 12, 2004 12:43 pm

Calling Session.lock(object, LockMode.READ) on a stale object prints an annoying StackTrace before throwing a StaleObjectStateException.

According to the API specs, this exception is expected so I believe Hibernate shouldn't print any stack trace when such event occurs.

We are using the Session.lock() to detect changes and invalidate our cached data. In this case, the exception is nothing that can occur and the stack trace in the logs is not meaningfull.

Does someone have a workaround for this ?

max · **Posted:** Mon Jan 12, 2004 1:31 pm

I find that to believe since hibernate don't do ex.printstacktrace() anywhere in the core.

Please show the file and linenumber in hibernate that should do this ? (you got the source ;)

bertrand · **Posted:** Mon Jan 12, 2004 4:26 pm

Hibernate doesn't print the stacktrace itself, but a call to the LOG api does ;-)

Here is the guilty code fragment (StaleObjectStateException constructor):

Code:

public StaleObjectStateException(Class persistentClass, Serializable identifier) {
      super("Row was updated or deleted by another transaction (or unsaved-value mapping was incorrect)");
      this.persistentClass = persistentClass;
      this.identifier = identifier;
      LogFactory.getLog(StaleObjectStateException.class).error("An operation failed due to stale data", this);
   }

BTW: I don't think it is good practice to call the LogFactory like this inside the Exception constructor itself...

max · **Posted:** Mon Jan 12, 2004 4:57 pm

hmmm...i agree it ain't beautifull....but it sure is nice to have it there in most circumstances....but i'll suggest you submit a request to the JIRA and explain why you don't like it.

bertrand · **Posted:** Wed Jan 14, 2004 5:33 am

http://opensource.atlassian.com/projects/hibernate/secure/ViewIssue.jspa?key=HB-624 has been created for this issue.