You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@commons.apache.org by "Ate Douma (JIRA)" <ji...@apache.org> on 2018/10/17 22:50:00 UTC

[jira] [Updated] (SCXML-285) Before executing invoke handlers after a macrostep all internal events must have been processed

     [ https://issues.apache.org/jira/browse/SCXML-285?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Ate Douma updated SCXML-285:
----------------------------
    Summary: Before executing invoke handlers after a macrostep all internal events must have been processed  (was: Before invoking invoke handlers after a macrostep all internal events must have been processed)

> Before executing invoke handlers after a macrostep all internal events must have been processed
> -----------------------------------------------------------------------------------------------
>
>                 Key: SCXML-285
>                 URL: https://issues.apache.org/jira/browse/SCXML-285
>             Project: Commons SCXML
>          Issue Type: Bug
>    Affects Versions: 2.0
>            Reporter: Ate Douma
>            Assignee: Ate Douma
>            Priority: Major
>             Fix For: 2.0
>
>
> I discovered a tricky bug in SCXMLSemanticsImpl#macroStep where it doesn't comply with the SCXML specification, part 3.13, which defines the following rules (among many others):
>  * A _macrostep_ is a series of one or more microsteps ending in a configuration where the internal event queue is empty [...]
>  * After completing a macrostep, the SCXML Processor _must_ execute in document order the <invoke> handlers [...]
> The tricky bug is that the current implementation does check for internal events first, but only processes the first event!
> In case of more than 1 internal event, the invoke handlers will be executed *before* (thereafter) processing any remaining internal events.
> The  [SCXML IRP Tests|https://www.w3.org/Voice/2013/scxml-irp/] didn't catch/test this condition, and it only in special scenarios will actually make a difference.
> I'll add a dedicated test for this, and then apply a fix proving it.  
>  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)