Multi-sequence evaluation

Jan Tluka jtluka at redhat.com
Tue Aug 14 15:47:29 UTC 2012


Tue, Aug 14, 2012 at 05:26:13PM CEST, jpirko at redhat.com wrote:
>Tue, Aug 14, 2012 at 05:11:56PM CEST, jtluka at 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
>>LNST-developers at lists.fedorahosted.org
>>https://lists.fedorahosted.org/mailman/listinfo/lnst-developers


More information about the LNST-developers mailing list