Multi-sequence evaluation

[Jan Tluka]

Hello everybody,

I'm thinking about changing current behavior in evaluating recipes with
more than one command sequence.

Currently if one of the command sequence fails the whole recipe is
aborted and the rest of the command sequences won't be executed.

Imagine you have the recipe split into several logical parts and you'd
like to see if any of them pass. Currently you have to rely on all of them 
to pass.

I propose to add 'dont_stop' or 'continue_on_fail' option to
command_sequence node. If it's set the following command sequence get's
executed.

e.g.
    <command_sequence continue_on_fail="yes">
        <command simple_test>
    </command_sequence>

    <command_sequence>
        <command simple_test_gets_executed_independent_on_previous_fail>
    </command_sequence>

    <command_sequence>
        <command simple_test_gets_executed_if_previous_pass>
    </command_sequence>

Feel free to share your ideas!

-Jan

[Jiri Pirko]

Tue, Aug 14, 2012 at 05:11:56PM CEST, jtluka(a)redhat.com wrote:
> Hello everybody,
>
> I'm thinking about changing current behavior in evaluating recipes with
> more than one command sequence.
>
> Currently if one of the command sequence fails the whole recipe is
> aborted and the rest of the command sequences won't be executed.
>
> Imagine you have the recipe split into several logical parts and you'd
> like to see if any of them pass. Currently you have to rely on all of them 
> to pass.
>
> I propose to add 'dont_stop' or 'continue_on_fail' option to
> command_sequence node. If it's set the following command sequence get's
> executed. 

I think we were talking about this already in past.

I think that it probably makes sense. How about "notfail" as default and
if user wants to specify some crucial sequence, add attribute
"quit_on_fail" ?

Jiri

> e.g.
>     <command_sequence continue_on_fail="yes">
>         <command simple_test>
>     </command_sequence>
>
>     <command_sequence>
>         <command simple_test_gets_executed_independent_on_previous_fail>
>     </command_sequence>
>
>     <command_sequence>
>         <command simple_test_gets_executed_if_previous_pass>
>     </command_sequence>
>
> Feel free to share your ideas!
>
> -Jan
> _______________________________________________
> LNST-developers mailing list
> LNST-developers(a)lists.fedorahosted.org
> https://lists.fedorahosted.org/mailman/listinfo/lnst-developers

[Jan Tluka]

Tue, Aug 14, 2012 at 05:26:13PM CEST, jpirko(a)redhat.com wrote:
> Tue, Aug 14, 2012 at 05:11:56PM CEST, jtluka(a)redhat.com wrote:
> >Hello everybody,
> >
> >I'm thinking about changing current behavior in evaluating recipes with
> >more than one command sequence.
> >
> >Currently if one of the command sequence fails the whole recipe is
> >aborted and the rest of the command sequences won't be executed.
> >
> >Imagine you have the recipe split into several logical parts and you'd
> >like to see if any of them pass. Currently you have to rely on all of them 
> >to pass.
> >
> >I propose to add 'dont_stop' or 'continue_on_fail' option to
> >command_sequence node. If it's set the following command sequence get's
> >executed.
>
>
> I think we were talking about this already in past.
>
> I think that it probably makes sense. How about "notfail" as default and
> if user wants to specify some crucial sequence, add attribute
> "quit_on_fail" ?
>
> Jiri
>  
Makes sense. I'll create ticket for this in trac. Also 'quit_on_fail'
sounds much better.

> >
> >e.g.
> >    <command_sequence continue_on_fail="yes">
> >        <command simple_test>
> >    </command_sequence>
> >
> >    <command_sequence>
> >        <command simple_test_gets_executed_independent_on_previous_fail>
> >    </command_sequence>
> >
> >    <command_sequence>
> >        <command simple_test_gets_executed_if_previous_pass>
> >    </command_sequence>
> >
> >Feel free to share your ideas!
> >
> >-Jan
> >_______________________________________________
> >LNST-developers mailing list
> &gt;LNST-developers(a)lists.fedorahosted.org
> >https://lists.fedorahosted.org/mailman/listinfo/lnst-developers

[Radek Pazdera]

On Tue, Aug 14, 2012 at 05:47:29PM +0200, Jan Tluka wrote:
>  Tue, Aug 14, 2012 at 05:26:13PM CEST, jpirko(a)redhat.com wrote:
>  >Tue, Aug 14, 2012 at 05:11:56PM CEST, jtluka(a)redhat.com wrote:
>  >>Hello everybody,
>  >>
>  >>I'm thinking about changing current behavior in evaluating recipes with
>  >>more than one command sequence.
>  >>
>  >>Currently if one of the command sequence fails the whole recipe is
>  >>aborted and the rest of the command sequences won't be executed.
>  >>
>  >>Imagine you have the recipe split into several logical parts and you'd
>  >>like to see if any of them pass. Currently you have to rely on all of them 
>  >>to pass.
>  >>
>  >>I propose to add 'dont_stop' or 'continue_on_fail' option to
>  >>command_sequence node. If it's set the following command sequence get's
>  >>executed.
>  >
>  >
>  >I think we were talking about this already in past.
>  >
>  >I think that it probably makes sense. How about "notfail" as default and
>  >if user wants to specify some crucial sequence, add attribute
>  >"quit_on_fail" ?
>  >
>  >Jiri
>  >
>  
>  Makes sense. I'll create ticket for this in trac. Also 'quit_on_fail'
>  sounds much better.
>  
Sounds good to me too :). I can have a look at it (it will require
a couple of changes in NetTestParse and NetTestController).

The attribute should be bolean 'quit_on_fail', set to zero by default?

>  >>
>  >>e.g.
>  >>    <command_sequence continue_on_fail="yes">
>  >>        <command simple_test>
>  >>    </command_sequence>
>  >>
>  >>    <command_sequence>
>  >>        <command simple_test_gets_executed_independent_on_previous_fail>
>  >>    </command_sequence>
>  >>
>  >>    <command_sequence>
>  >>        <command simple_test_gets_executed_if_previous_pass>
>  >>    </command_sequence>
>  >>
>  >>Feel free to share your ideas!
>  >>
>  >>-Jan
>  >>_______________________________________________
>  >>LNST-developers mailing list
>  &gt;&gt;LNST-developers(a)lists.fedorahosted.org
>  >>https://lists.fedorahosted.org/mailman/listinfo/lnst-developers
>  _______________________________________________
>  LNST-developers mailing list
>  LNST-developers(a)lists.fedorahosted.org
>  https://lists.fedorahosted.org/mailman/listinfo/lnst-developers

[Jan Tluka]

Wed, Aug 15, 2012 at 02:29:15PM CEST, rpazdera(a)redhat.com wrote:
> On Tue, Aug 14, 2012 at 05:47:29PM +0200, Jan Tluka wrote:
> > Tue, Aug 14, 2012 at 05:26:13PM CEST, jpirko(a)redhat.com wrote:
> > >Tue, Aug 14, 2012 at 05:11:56PM CEST, jtluka(a)redhat.com wrote:
> > >>Hello everybody,
> > >>
> > >>I'm thinking about changing current behavior in evaluating recipes with
> > >>more than one command sequence.
> > >>
> > >>Currently if one of the command sequence fails the whole recipe is
> > >>aborted and the rest of the command sequences won't be executed.
> > >>
> > >>Imagine you have the recipe split into several logical parts and you'd
> > >>like to see if any of them pass. Currently you have to rely on all of them 
> > >>to pass.
> > >>
> > >>I propose to add 'dont_stop' or 'continue_on_fail' option to
> > >>command_sequence node. If it's set the following command sequence
> get's
> > >>executed.
> > >
> > >
> > >I think we were talking about this already in past.
> > >
> > >I think that it probably makes sense. How about "notfail" as default
> and
> > >if user wants to specify some crucial sequence, add attribute
> > >"quit_on_fail" ?
> > >
> > >Jiri
> > >
> > 
> > Makes sense. I'll create ticket for this in trac. Also 'quit_on_fail'
> > sounds much better.
> >
>
> Sounds good to me too :). I can have a look at it (it will require
> a couple of changes in NetTestParse and NetTestController).
>
> The attribute should be bolean 'quit_on_fail', set to zero by default?
>  
I have the work already done. Should be sent in few minutes.

-Jan

> > >>
> > >>e.g.
> > >>    <command_sequence continue_on_fail="yes">
> > >>        <command simple_test>
> > >>    </command_sequence>
> > >>
> > >>    <command_sequence>
> > >>        <command
> simple_test_gets_executed_independent_on_previous_fail>
> > >>    </command_sequence>
> > >>
> > >>    <command_sequence>
> > >>        <command simple_test_gets_executed_if_previous_pass>
> > >>    </command_sequence>
> > >>
> > >>Feel free to share your ideas!
> > >>
> > >>-Jan
> > >>_______________________________________________
> > >>LNST-developers mailing list
> > &gt;&gt;LNST-developers(a)lists.fedorahosted.org
> > >>https://lists.fedorahosted.org/mailman/listinfo/lnst-developers
> > _______________________________________________
> > LNST-developers mailing list
> > LNST-developers(a)lists.fedorahosted.org
> > https://lists.fedorahosted.org/mailman/listinfo/lnst-developers

[Radek Pazdera]

On Wed, Aug 15, 2012 at 02:45:17PM +0200, Jan Tluka wrote:
>  Wed, Aug 15, 2012 at 02:29:15PM CEST, rpazdera(a)redhat.com wrote:
>  >On Tue, Aug 14, 2012 at 05:47:29PM +0200, Jan Tluka wrote:
>  >> Tue, Aug 14, 2012 at 05:26:13PM CEST, jpirko(a)redhat.com wrote:
>  >> >Tue, Aug 14, 2012 at 05:11:56PM CEST, jtluka(a)redhat.com wrote:
>  >> >>Hello everybody,
>  >> >>
>  >> >>I'm thinking about changing current behavior in evaluating recipes
> with
>  >> >>more than one command sequence.
>  >> >>
>  >> >>Currently if one of the command sequence fails the whole recipe is
>  >> >>aborted and the rest of the command sequences won't be executed.
>  >> >>
>  >> >>Imagine you have the recipe split into several logical parts and
> you'd
>  >> >>like to see if any of them pass. Currently you have to rely on all of
> them 
>  >> >>to pass.
>  >> >>
>  >> >>I propose to add 'dont_stop' or 'continue_on_fail'
> option to
>  >> >>command_sequence node. If it's set the following command sequence
> get's
>  >> >>executed.
>  >> >
>  >> >
>  >> >I think we were talking about this already in past.
>  >> >
>  >> >I think that it probably makes sense. How about "notfail" as
> default and
>  >> >if user wants to specify some crucial sequence, add attribute
>  >> >"quit_on_fail" ?
>  >> >
>  >> >Jiri
>  >> >
>  >> 
>  >> Makes sense. I'll create ticket for this in trac. Also
> 'quit_on_fail'
>  >> sounds much better.
>  >>
>  >
>  >Sounds good to me too :). I can have a look at it (it will require
>  >a couple of changes in NetTestParse and NetTestController).
>  >
>  >The attribute should be bolean 'quit_on_fail', set to zero by default?
>  >
>  
>  I have the work already done. Should be sent in few minutes.
>  
>  -Jan 
Oh, ok, that's great! :)

I'll take something else then.

>  
>  >> >>
>  >> >>e.g.
>  >> >>    <command_sequence continue_on_fail="yes">
>  >> >>        <command simple_test>
>  >> >>    </command_sequence>
>  >> >>
>  >> >>    <command_sequence>
>  >> >>        <command
> simple_test_gets_executed_independent_on_previous_fail>
>  >> >>    </command_sequence>
>  >> >>
>  >> >>    <command_sequence>
>  >> >>        <command simple_test_gets_executed_if_previous_pass>
>  >> >>    </command_sequence>
>  >> >>
>  >> >>Feel free to share your ideas!
>  >> >>
>  >> >>-Jan
>  >> >>_______________________________________________
>  >> >>LNST-developers mailing list
>  >> &gt;&gt;LNST-developers(a)lists.fedorahosted.org
>  >> >>https://lists.fedorahosted.org/mailman/listinfo/lnst-developers
>  >> _______________________________________________
>  >> LNST-developers mailing list
>  >> LNST-developers(a)lists.fedorahosted.org
>  >> https://lists.fedorahosted.org/mailman/listinfo/lnst-developers