In response to yesterday's ponderings it was quickly pointed out that I could easily run two agents in a Form's WQS event by simply calling one from within the other. Imagine this code in the first agent:

Dim agentTwo As NotesAgent
Set agentTwo = database.GetAgent("(wqs Agent Two)")

agentTwo.Run( document.NoteID )

All seems fairly obvious and I don't know why I didn't think to try it in the first place.

However, there's a problem with this approach. Running the second agent against the document in which the first WQS agent was triggered only works if you save the document before running Agent Two and then again before leaving Agent Two.

The first save within Agent One is needed for two reasons.

1. A new document doesn't have a NoteID, which you need to pass to the other agent, until it's been saved.
2. Any changes you make to the document in the code in Agent One won't be reflected in the document you see in Agent Two. This is because Agent Two is getting a handle on the document that has been committed to the disk and not the context document, which Agent One is working on.

The second save is called at the end of Agent Two and is needed because we're operating outside the context of the document being saved. We won't be able to access the altered values in Agent One after Agent Two has run as we're effectively working on different instances of the same document. At least that's how I understand it. Although in practice I found that changes made in Agent Two don't commit anyway as they're over-written by the impending save of the document from the completion of Agent One, which is the controlling WQS agent.

All in all, while testing this, I found it doesn't really pan out like you'd hope it would and it's not an approach I'd recommend or like to use myself. The only time I'd use it is if the code in Agent Two was merely reading the document and not making any changes. Then there's the matter of calling doc.save() from within a WQS agent, which I never like to do. It just seems wrong.

In summary, if you want to run two agents in the WQS event -- one LotusScript, one Java -- then I'd still recommend running them one after the other, as shown yesterday and avoid having to save a document which is already being saved. If you want to run Java that merely needs read access to the document then it might work for you.

If you're interested in seeing how I came to the above conclusions here's the database I was messing about with. Put it on a server and watch the console while saving documents. You should get an idea of what's going on.

I'm still leaning toward re-writing the agent in Java though...