[Linaro-validation] lava-deployment-tool

Andy Doan andy.doan at canonical.com
Wed Feb 15 18:15:32 UTC 2012


On 02/15/2012 12:00 PM, Zygmunt Krynicki wrote:
> Hi Andy
>
> Not using l-d-t is very much possible _but_ as we grow it will be
> harder and less likely to work. A plain virtualenv will obviously work
> (l-d-t also sets one up) as long as you don't care about performance
> and don't plan to use the scheduler.
>

For me, when I start with something new I want something that's not 
going mess with my computer in ways I don't yet understand. I think it 
would be a real plus and goal to keep things in a way that this type of 
model can be used. There are obviously things like the scheduler where 
that probably can't be avoided. However, I'm guessing a number of LAVA 
consumers are going to be like me. They want to create a test and 
possibly tweak a view of the data.

I think there might be 4 paths to help guys in my situation:

  1. The approach I've been doing which may not work down the road

  2. Update l-d-t to support some type of "develop" mode which does
     what 1 attempts

  3. A new script lava-develop-tool. Its like 2 but splits things out
     into a new script. Just suggesting this in the event people like
     2, but don't think its maintainable.

  4. Tell me to stop being anal-retentive, grow up, and use the tool.

I can see good arguments for all options :)

-andy



More information about the linaro-validation mailing list