James Lent
unread,Oct 25, 2012, 11:53:40 AM10/25/12Sign in to reply to author
Sign in to forward
You do not have permission to delete messages in this group
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to hamake...@googlegroups.com
I am very interested in using HAMAKE for a project I am working on. The general architecture fits my needs very well.
In downloads the jar is version: 20b-3. That seems to have been built in June 1010 based on r331 out of trunk. There have been many updates since then so I am also considering using the version that can be built from trunk (20b-4). Does anyone have any recommendations?
In my initial testing I have noticed some difference in behavior between the two. In one case I prefer the 20b-3 behavior and in one case I prefer the 20b-4 behavior. This is making the decision a little more difficult for me.
I have a simple linear set of tasks A-->B-->C-->D dataflow defined.
If I change the data coming into A, remove the output of D, and break Task A then the behavior I see is:
20b-3: Task A fails (as expected), but, Tasks B, C, and D try to run and D actually does?!?!?!?!
20b-4: Task A fails. Tasks B, C, and D are removed from the graph (e.g., don't run) as soon as A fails
If I just change the data coming into Task A:
20b-3: Tasks A, B, C, and D all rerun
20b-4: Task A fails because its output data already exists?!?!?!?!
It is possible my dataflow definition is incorrect, but, it does yield my expected behavior with one of the two jars in each case.