RE: Run global <script> before or after <datamodel> initialization?

F.Y.I, here's what I think the start-up code should look like.  Perhaps 'running' should be set to 'true' earlier ( because you clearly are running, in some sense, when you initialize the datamodel and execute  the global script element), but it is used only in the main event loop.

procedure interpret(doc):
    if not valid(doc): failWithError()
    expandScxmlSource(doc)
    configuration = new OrderedSet()
    statesToInvoke = new OrderedSet()
    internalQueue = new Queue()
    externalQueue = new BlockingQueue()
    historyValue = new HashTable()
    datamodel = new Datamodel(doc)
    if doc.binding == "early":
        initializeDatamodel(datamodel, doc)
    executeGlobalScriptElement(doc)
    enterStates([doc.initial.transition])
    running = true
    mainEventLoop()


-          Jim

From: Gavin Kistner [mailto:phrogz@me.com]
Sent: Thursday, February 14, 2013 4:43 PM
To: www-voice@w3.org; www-voice@w3.org
Subject: Run global <script> before or after <datamodel> initialization?

The prose in the official algorithm describes a <script> element at the root of the document as being run after the data model is initialized:

"Create an empty configuration complete with a new populated instance of the data model and a[sic] execute the global scripts[sic]."

(I assume that "scripts" should be "script", as only one <script> element may appear as a child of <scxml>.)

This order makes sense to me: I want my script to be able to read initial datamodel values and manipulate them. (I'm assuming that script executes in the context of the data model and may thus set values in it as well.)

However, the official pseudo-code executes the global script element before the data model is initialized:

    datamodel = new Datamodel(doc)
    executeGlobalScriptElements(doc)  # Note: this is also plural
    ...
    if binding == "early":
      initializeDatamodel(datamodel, doc)

Is there a compelling reason to execute global <script> before initializing the data model?

Received on Thursday, 14 February 2013 23:27:39 UTC