Rerunning failed scenario inside Around block

[James Poyser]

I'd like to find a way to re-run a failed scenario, before it's reported back (to our CI server in our case). If it fails twice, then it's reported as a failure.

This is to improve the reliability of browser automation tests. I'd like to set a global variable to slow down execution time on the second run. If it still fails, then it's a 'real' failure.

I was hoping this would work, but alas not - any ideas?

Around do |scenario, block|

  begin

    $slowdown = 0

    block.call

  rescue

    $slowdown = 1

    block.call

  end

end

[Paolo Ambrosio]

Can't help you with the Ruby part but I really hope you won't need it
after reading my comments below.

On Sun, Feb 7, 2016 at 9:58 AM, James Poyser <ja...@inniaccounts.co.uk> wrote:

> I'd like to find a way to re-run a failed scenario, before it's reported
> back (to our CI server in our case). If it fails twice, then it's reported
> as a failure.

Instead of trying this hack, what about fixing the root cause?

> This is to improve the reliability of browser automation tests. I'd like to
> set a global variable to slow down execution time on the second run. If it
> still fails, then it's a 'real' failure.

In general you should poll for some condition to be met instead of
introducing slowdowns. This way your tests will be as fast as your
system can be.

[James Poyser]

It's an admirable aim Paolo, however this simply doesn't work in the real world of browser automation, where there's inherence flakiness and inconsistencies in the frameworks which go beyond our control.

[Tim Walker]

On Feb 8, 2016 18:17, "James Poyser" <ja...@inniaccounts.co.uk> wrote:
>
> It's an admirable aim Paolo, however this simply doesn't work in the real world of browser automation, where there's inherence flakiness and inconsistencies in the frameworks which go beyond our control.
>

When determinism is lost smash the state...

>
> On Sunday, 7 February 2016 14:44:28 UTC, Paolo Ambrosio wrote:
>>
>> Can't help you with the Ruby part but I really hope you won't need it
>> after reading my comments below.
>>
>> On Sun, Feb 7, 2016 at 9:58 AM, James Poyser <ja...@inniaccounts.co.uk> wrote:
>>
>> > I'd like to find a way to re-run a failed scenario, before it's reported
>> > back (to our CI server in our case). If it fails twice, then it's reported
>> > as a failure.
>>
>> Instead of trying this hack, what about fixing the root cause?
>>
>> > This is to improve the reliability of browser automation tests. I'd like to
>> > set a global variable to slow down execution time on the second run. If it
>> > still fails, then it's a 'real' failure.
>>
>> In general you should poll for some condition to be met instead of
>> introducing slowdowns. This way your tests will be as fast as your
>> system can be.
>>
>> > I was hoping this would work, but alas not - any ideas?
>> >
>> > Around do |scenario, block|
>> >   begin
>> >     $slowdown = 0
>> >     block.call
>> >   rescue
>> >     $slowdown = 1
>> >     block.call
>> >   end
>> > end
>

> --
> Posting rules: http://cukes.info/posting-rules.html
> ---
> You received this message because you are subscribed to the Google Groups "Cukes" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to cukes+un...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

[Samar Mukherjee]

Can anyone help on the Cucumber Java to rerun the failed scenario?

[Matt Wynne]

https://github.com/cucumber/cucumber-ruby/issues/882

Please add feedback / help there.