Remove early optimization.

Just to trigger tests.

  This implementation is not optimal from performance point of view,
  however the fix is more urgent and the current implementation has
  nearly the same performance outcome.

  Optimisation should come later on query less objects per time.

Increase revision of BT.

Also return created objects for _createComputer.

getUrlString will be used to link the Computer to a External Inventory and/or Documentation System.

  Don't show close computers on rest API when list, it is not relevant for monitoring
  Don't list request_destroyed software instances or looking for partition, as it is not
    relevent for monitoring.

Refactor and rename RSS Listbox implemented for Support Request
Include RSS action on Regularisation Module

Those fields are not used anymore.

Only generate a limited number of Support Requests per time (For now hardcoded in 3).
Drop duplicated cases when computer is not contacting already.
Don't generate SR for recently created instances.

an absolute mock module is already used.

This is duplicated check with "Check if Computer Contacted". The only ocasion that this can happens is if someone comments "slapos node instance" from cron (in this case is not master duty to know if backend is not modified in a bad way).

This is a login feature (experimental and prototype) which permit login on SlapOS master using external logins (google, facebook, ...) accessing a url. SlapOS master works  on this case as bridge only, for setup cookies on master domain to permit usage o of external websites later.

Patch Computer_applyComputerModel in order to apply the capacity_quantity when apply the Computer Model.

Use Computer Model Capacity and Update Computer Capacity when discover capacity (else keep using previous maximum default value).

This change only affects Public Computers which are able to allocate (contacted recently the master).

Capacity can be set on Computer Models and be eventually applied on Computers

This change aims prevent flood of request and looking for partitions ghosts when hosting has full capacity or allocation is not possible.