Bugzilla – Bug 5583
CheckpointAlgorithm needs to specify timeunit (seconds) and other javadoc fixes
Last modified: 2015-09-03 17:42:26 UTC
int checkpointTimeout() throws Exception
does not specify the timeunit (seconds) for the return value in the current javadoc:
Its javadoc contains a @param, but this method has no param.
Just looking at this again. I ran Checkstyle against the API package, and got 157 warnings. A good chunk of these reflect the fact that we didn't comment each of the exception constructors, which I don't think is a big deal.But I'm going to have to spend some time sorting through this.
Fixed this one method with this Javadoc text:
* The checkpointTimeout is invoked at the beginning of a new
* checkpoint interval for the purpose of establishing the checkpoint
* It is invoked before the next checkpoint transaction begins. This
* method returns an integer value, which is the timeout value (expressed
* in seconds) which will be used for the next checkpoint transaction.
* This method is useful to automate the setting of the
* checkpoint timeout based on factors known outside the job
* @return the timeout interval (time unit = seconds) to use for the
* next checkpoint interval
* @throws Exception thrown for any errors.
public int checkpointTimeout() throws Exception;
Made some other changes.. I don't think it's too interesting to list them all. A couple were bugs but most were additions of things that had been omitted up until now.
One I caught was: in the spec we have a method described as StepExecution#getUserPersistentData but in the RI and TCK it's been changed to StepExecution#getPersistentUserData. So I updated the spec accordingly (I may have mentioned this in another bug).
Note the spec contains a lot of the Javadoc but not 100% of the RI's API Javadoc, so not every update will go into the spec. Will try to refresh the spec with the latest Javadoc but this is a bit of a manual merge and my master copy is the RI.
In other words, help keep me honest here if you see additional bugs in Javadoc.