Support partitioning at the scenario level instead of feature file level

[GoogleCodeExporter]

Hi

Right now, I TLB considers a feature file as the smallest unit. In other words, 
if we have one feature file with 10 scenario and if we have 2 partitions, this 
feature file will be passed to one balancer and other will be idle.
Another example is , if we have two feature file , one with 15 scenarios and 
other with 2, we will see a little benefit from TLB unless both feature file 
takes almost the same time.

Is there any plan to add a feature where individual scenarios can be passed to 
balancers instead of feature files ?

This is neither a big problem right now nor a blocking issue, but i think a 
nice to have feature. I can see this will have performance impact on TLB 
because of parsing the feature file, scenarios which relies on background 
keyword,etc.

Regards
Ragavan

Original issue reported on code.google.com by singh.janmejay on 30 Jun 2012 at 6:15

[GoogleCodeExporter]

I am running into the same thing. This would be great. 

Original comment by JoeNoona...@gmail.com on 1 Apr 2014 at 3:56

[GoogleCodeExporter]

@Pavan: doable? 

Original comment by singh.janmejay on 2 Apr 2014 at 4:14

[GoogleCodeExporter]

When this was implemented, Cucumber let us give it a list of feature files and 
did not let us pass just a list of scenarios. This was a limitation in the 
filtering capability of cucumber's runner. 

From TLB's point of view, we can very easily change to scenario level filtering 
as soon as Cucumber supports this. 

One point to note is if TLB starts supporting this, your test run history might 
be rendered useless since feature level values will be no longer useful and you 
will start from scratch.

Original comment by itspa...@gmail.com on 2 Apr 2014 at 4:21

[GoogleCodeExporter]

Can anyone confirm if cucumber now has support for scenario level
selection? If not we'll have to investigate.

Original comment by singh.janmejay on 2 Apr 2014 at 12:40