Sebastien Petitdemange authored Jul 03, 2017 and Emmanuel Papillon committed Jul 03, 2017

Actually musst card produce signed data

Sebastien Petitdemange authored Jun 09, 2017 and Emmanuel Papillon committed Jun 09, 2017

It's better to wait the end of all stop command.
Previously exit the joinall at the first exception seen.

Sebastien Petitdemange authored Jun 07, 2017 and Emmanuel Papillon committed Jun 07, 2017

Fixing a problem on some controller which has a closed loop...
Time to time at the end of the movement, it send READY then MOVING
because it reactivate the closed loop. So if no luck axis read MOVING and
stay in this state forever.

Actually the last state was never send because _hw_control was False

why do we need to have complex task decorator?

wait motion in parallel, because if an exception occurred in one of
the axis re-raise the exception immediately (don't wait other axis)

return the move task status

modify a little but the wait_move to be atomic.

The communication is now managed with bliss tcp in
function _command and _askcommand in file system.py

changed also to default logging python module

Tcp command fixes

Closes #90

See merge request !391

Motion fixes

See merge request !392

Some exception was hidden because of the wait in sequence.

Previously, standard movement only managed KeyboardInterrupt exception.
This was wrong because the task is killed with an Greenlet exit.

close issue #90

avoid call to redis when testing if object is a MeasurementGroup instance

See merge request !388

Fixes

See merge request !387

New `check_encoder` config parameter for Axis object

See merge request !386

Fix issue 72

Closes #72

See merge request !385

set to False. When set to True, if an encoder is present, at the end
of motion it is doing encoder reading to compare controller position with
encoder position - an exception is raised if the two positions are not
within the tolerance