After 1.6.1 was deployed we have two processes, stop the redundant process.

In 58fbaabf (svcbackend/standalone: use shorter names for supervisor sockets
, 2020-07-17) we changed the socket path from supervisord.sock to a shorter
sv.sock, but this caused issues when updating running slapos node instances,
because this was trying to start a new supervisord process using the new
socket path, while the old one was still running.

To keep services running after upgrade, introduce a compatibility layer in
this socket path - if we still have a supervisord running with a socket at
this previous path, keep using this old socket name.

Output is already captured on SlapPopen instances and the output
is available with output attribute.

    User can view his Cloud Contract and its consumption.

      This is used to Sales website when a Invitation token is provided and it will be used when request button will be available on logged user website.

    This consider Maximum Delay and Maximum Credit a user can have

   This is a faster version to later accomodate custom implementation from Cloud Contracts

    This help us to avoid query outdated catalog, consider that the user is only at transaction for now.

  This requires more work before continue.

See merge request nexedi/slapos.core!245

When test machine is busy it happens that running a simple promise takes more
than one second and this cause some test to error with "Promise timed out after
1 seconds" in cases where the promise is not supposed to have a timeout, it's
just that sometimes it takes more than the default timeout of 1 second.

   Cover the the update the update of allocation_scope
   Cover the Computer Network set Reference
   Cover the bootstrap of Project
   Cover setRole on Organisation

Helper function to return the configuration used by a promise plugin.

Helper class to check crontabs in software release tests

    This makes a direct edit w/o introspect previous value.

   We don't need to generate password everywhere