[LRUG] The August Meeting

[Matt Wynne]

That was and still is our approach at Songkick, but we have to be  
pretty gentle with the build with 8 developers committing to it, so  
anything that gives me an early warning if I've borked it is welcome.  
Unfortunately running all 7000+ cuke steps locally isn't really an  
option any more. Even I can't drink that much tea.

On 20 Jul 2009, at 15:21, Matthew Rudy Jacobs wrote:

> My approach is;
>
> - run the tests I think will have changed locally
> - run all the tests every time on CI
>
> B00M!
>
> 2009/7/20 Matt Wynne <matt at mattwynne.net>
>
> On 17 Jul 2009, at 12:03, Paul Battley wrote:
>
> 2009/7/17 Matt Wynne <matt at mattwynne.net>:
> I'd quite like to spread the word about my lazy Cucumber runner:
> http://github.com/mattwynne/cucover/tree/master
>
> That's a great idea. I just tried it out. Unfortunately, it seems to
> make everything run so slowly that I'd be better off just running all
> the features ...
>
> Paul.
>
> Yes... unfortunately it's more of a proof of concept right now - I  
> don't use it myself day-to-day as too much of our stuff runs out of  
> process for me to be able to use it reliably. But I had noticed  
> about the speed - I'm not sure how much that's unavoidable RCov  
> overhead and how much is my naive first-pass implementation which is  
> chattering away to disk all the time.
>
> If you have a significant number of features though, focussed  
> regression runs are a big time-saver, especially when you're  
> refactoring and lose track of what you might have broken. That's the  
> theory anyway... :)
>
>
> cheers,
> Matt Wynne
>
> http://mattwynne.net
> +447974 430184
>
> _______________________________________________
> Chat mailing list
> Chat at lists.lrug.org
> http://lists.lrug.org/listinfo.cgi/chat-lrug.org
>
> _______________________________________________
> Chat mailing list
> Chat at lists.lrug.org
> http://lists.lrug.org/listinfo.cgi/chat-lrug.org

cheers,
Matt Wynne

http://mattwynne.net
+447974 430184