Extension is attempting to restart docker for every playbook run

[zikalino]

In addition second execution fails, because extension tries to run "docker run" within running docker terminal. I think running docker again should be avoided, as user may be using terminal to install something, for example ansible role, and that will be gone if docker is restarted. So we should only restart docker when really necessary. Also, i think there should be some option to commit any new changes user made in docker image? or that should be left to the user to do manually?

[yaweiw]

think the latest behavior is F1 run in terminal will launch a new terminal where original one stays intact. Max limit of active terminals is 20 if memory serves.

[zikalino]

i think we should have more thorough discussion on this, as i mentioned, in most cases user will want to reuse the same terminal as long as possible.

[yungez]

user would change credential file, playbook file path etc. comparing to guess what user would like do, it's more clean to create a new terminal , while we could see if it's possible to reuse container instance if new terminal. terminal new each time, but contains keep using same.

[zikalino]

we should consider following screnario:

• start the terminal
• install ansible role
• run the playbook and role is gone because docker container was restarted

i think the simplest solution would be just restarting the terminal when user closes old terminal. so if something changes we could advise user just to close the terminal.

i like some cotinuity when using terminal, for example remembering my past commands, etc.

[zikalino]

wrt playbook file path:

• i think the best assumption would be that playbook is in a current workspace, and just map current workspace
• credentials could be passed to running docker container in different way, or we can spot that they were indeed changed

i would still insist that the terminal should be restarted only if we detect it's necessary, or when closed manually by user.