What kind of process should we have to decide what drivers to adopt / give repos?

machine-drivers / discussion

Lets get started

Apache License 2.0

1 stars 0 forks source link

What kind of process should we have to decide what drivers to adopt / give repos? #1

Open SvenDowideit opened 6 years ago

SvenDowideit commented 6 years ago

[ ] do we need to be selective?
[ ] are there any open source drivers we want/need to not accept, or track in this org?
[ ] ... please add other questions :)

SvenDowideit commented 6 years ago

IMO (separate from creating this issue), is that we should collect all machine drivers here, even if they're just mirrors of upstream drivers. That would enable users to have one common collation point - and for mirror repos, we'd turn off issues&pr's - and possilby automate syncing.

the utterly distributed nature of the drivers atm is both awesome wrt getting things done, and terrible wrt finding things, and being able to fix larger scale problems in the code.

r2d4 commented 6 years ago

Is there an testing plan in place yet?

I think that whatever drivers that are here need need to be tested. The testing infrastructure has historically been difficult, since there are a few different OSes and each requires virtualization to be enabled. In minikube, we've solved this by a combination of physical machines as well as VMs with nested virtualization enabled. However, the cost of maintenance is high.

afbjorklund commented 6 years ago

Don't have any automated tests in place yet, so just experimenting on users (including myself)

This was for qemu (and kvm/kvm2). Using virtualbox as the reference driver, in case of doubt...