This issue is a placeholder for comments on this blog post which is directing folks to add their comments here in the interest of hosting the follow-up discussion somewhere within the environment.


jthorson’s picture

Issue summary: View changes

Updated issue summary.

rfay’s picture

Exceptional summary of the risks.

In a perfect world I would definitely prefer to start over, using what we have learned in the real world to simplify the testing system and make it more flexible.

But we'd have to have a pretty big energy commitment and a clear community mandate to take this risk. Otherwise, we can continue on the path of improving as we go. Jthorson's work over the last 9 months has made approach far more likely to succeed.

boombatower’s picture

As jthorson and I discussed the plan moving forward will be to use the ReviewDriven codebase which has been rewritten with many of these things in mind.

rfay’s picture

@boombatower, I assume you were saying "I proposed to jthorson a plan that would involve using ReviewDriven". It would be good for you to expound on your plan here. I have heard it involves open-sourcing the code and running ReviewDriven on infrastructure (Yay!) ?

jthorson’s picture

I agree that the ReviewDriven architecture certainly looks promising, and it appears that it would address many of the goals identified in my post ... and I think it's definately worth adding it to the list of things to look at during the testbot sprint.

Whether we go with further enhancements on the existing modules, a 'fresh start' D7 build, implementing the ReviewDriven code, Jenkins branch testing, or something completely different ... one thing all these have in common is a need to do some significant UX work on the d.o/PIFT side of things. I also anticipate that we'll want to migrate the balance of control from qa.d.o to the d.o side of things, which would give us the environment decoupling and testing granularity that I'd like to see.

And, since it's been my personal crusade for such a long time, I'd also throw 'enable sandbox testing' out there as yet another thrust for the testbot sprint ... the blockers on this are on the project_issue side of things, and would also be a pre-requisite for all of the proposed evolutionary paths on the testbot side of things.

jthorson’s picture

Issue summary: View changes

Updated link

geerlingguy’s picture

Adding related issue.

Mixologic’s picture

Status: Active » Closed (outdated)

Drupalci is a thing. Things have evolved now.