W3C home > Mailing lists > Public > www-voice@w3.org > July to September 2014

Re: Preemption in Test 403c

From: Jim Barnett <1jhbarnett@gmail.com>
Date: Thu, 31 Jul 2014 12:01:35 -0400
Message-ID: <53DA685F.9010902@gmail.com>
To: Markus Weiland <markus@tracktik.com>, www-voice@w3.org
Yes.   conf:targetfail="" sets the target of the transition to be the 
<conf:fail> state.

  (The .txml files are meant to be transformed into scxml by means of an 
XSLT style sheet.  We provide sample ones for the javascript and xpath 
data models.  If you look at those transformations, you'll see that 
<conf:fail> gets converted into a <final> state with id="fail" and 
conf:targetfail="" gets converted into target="fail".  The idea behind 
this abstraction is to let platforms tweak the tests to suit their 
environments.  For example, you might want to write pass/fail results to 
a DB.  You can do this by editing the .xslt file rather than having to 
edit each test by hand.)

On 7/31/2014 11:50 AM, Markus Weiland wrote:
> Hi Jim,
> Thanks for your explanation. Much clearer now.
> So just to confirm, the transition attribute ``conf:targetfail=""`` in 
> the event1 transition of p0s3 is considered to specify a transition 
> out of s0 (presumably into ``conf:fail``)?
> On Thu, Jul 31, 2014 at 10:42 AM, Jim Barnett <1jhbarnett@gmail.com 
> <mailto:1jhbarnett@gmail.com>> wrote:
>     The transition in P0s4 is a targetless transition, and thus has an
>     empty exit set and does not conflict with any other transition.
>     Therefore it is never preempted.   For the same reason, the
>     transition in p0s1 is not preempted.
>     The transition for event1 in p0s3 has a non-empty exit set (namely
>     s0 and all its children), so it conflicts with and is preempted by
>     the transition in p0s2.  For event2 the transtion in p0s3 also
>     conflicts with the transition in p0s2, but in this case it does
>     the preempting (since transitions in descendents preempt
>     transitions in ancestors.)
>     On 7/30/2014 10:09 PM, Markus Weiland wrote:
>         Hi,
>         Can someone please explain, in test case 403c, why the
>         transition for "event1" is supposed to be preempted by p0s2
>         for the first transition in p0s3, but is not supposed to be
>         preempted for the same "event1" for the catchall transition in
>         p0s4?
>         In other words, what makes the transition in p0s3 different
>         from the transition in p0s4 so that it gets preempted? Also,
>         for the sake of understanding, is the transition for "event1"
>         in p0s1 preempted?
>         Thank you
>     -- 
>     Jim Barnett
>     Genesys

Jim Barnett
Received on Thursday, 31 July 2014 16:02:08 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:07:45 UTC