Possible Gscreen bug

When I use the standard gscreen restart dialog, the following error pops up:
*******************************************************
1
Traceback (most recent call last):
File "/usr/bin/gscreen", line 3793, in restart_dialog_return
self.widgets.gcode_view.set_line_number(value)
File "/usr/lib/python2.7/dist-packages/gladevcp/hal_sourceview.py", line 167, in set_line_number
self.highlight_line(self.gstat, self.offset)
File "/usr/lib/python2.7/dist-packages/gladevcp/hal_sourceview.py", line 182, in highlight_line
line = self.buf.get_iter_at_line(l-1)
TypeError: integer argument expected, got float

*******************************************************
If I use a same named method in the handler file but change gscreen line 3791 from:
value = calc.get_value()
to:
value = int(calc.get_value())
then everything works fine. This same error occurs even when running the sim.gscreen that comes with linuxcnc so it probably has nothing to do with my handler file. Of course, you only see this if you start linuxcnc from a terminal.
Using debian stretch, linuxcnc v.2.8 installed (not RIP)

Thank you I will look at this as soon as I can.

Chris M

fixed in 2.7 and master - thanks!

Chris M

To recycle an old topic, does gscreen support multiple joints per axis? I don't see anything about a joint_axis dictionary. My machine has 2 motors for the Y axis and now won't work with the 7i96 driver board.

I would bet there are some bugs lingering... I see there is no gantry sample config.
Can you describe the problem?

In my lathe config I haven't connected the joint jogging pins and wonder why it won't jog before it's homed - until I remember master is different now.

Chris M

Basically, in the INI file I have:
[TRAJ]
COORDINATES = X Y Y Z
...
[KINS]
KINEMATICS = trivkins coordinates=xyyz
JOINTS = 4

Gscreens self.data.axis_list is [x,y,z] which is probably OK but there's no mapping from axis letters to joints. HALSHOW does indicate that there are 4 joints but something is not right. When I try to home or jog the Z axis, I get :
joint 3 following error
emc/task/taskintf.cc 942: Error on joint 3, command number 278
This is likely because the code (handler file) is not using the right joint number in the jog command. I could hard code it but would rather have it work the way it's supposed to.

Are you using Silverdragon?

Yes, using silverdragon. However, further developments. I did some experiments with gmoccapy and axis and had the same problems. I finally got a XYZY configuration to work with all UIs including gscreen, at least in simulation mode. It turns out it's all in which HAL files are specified in the INI file. In the hallib directory of the sim config directories, there is a gantrysim.hal which must be incorporated. I basically merged core_sim.hal and gantrysim.hal into silverdragon.hal and now it works in simulation mode with gscreen. The next step is to move this to the real hardware and figure out how to make it work with the 7i96. This problem didn't appear when I used the parport and BOB because stepconf has an option to slave 2 axes together, while the 7i96 config tool does not.

Interesting -I'll try to find time to look into it.
I did find that xyyz and xyzy worked differently - I didn't figure out why - thanks for the update.

Chris M

My machine is now working with the 7i96 in XYZY configuration. I will try XYYZ eventually, but there is a new problem. *Sorry*
After updating to linuxcnc 2.9.0, gscreen now spits out the entire contents of some module, probably the handler function. Here's a partial listing:

**** GSCREEN INFO: No Screen 2 glade file present
adding import dir /home/jim/linuxcnc/configs/sim.gscreen.gscreen_custom
module 'silverdragon_handler' imported OK
module 'silverdragon_handler' : 'get_handlers' function found

task: main loop took 0.043250 seconds
task: main loop took 0.033703 seconds
Auto Z Reference - Enabled
task: main loop took 0.011900 seconds
task: main loop took 0.041333 seconds
timeout 100
**** GSCREEN INFO: postgui filename: silverdragon_postgui.hal

This doesn't happen with gmoccapy or axis. I believe it occurs at or around line 399 in gscreen.py.
Not sure if all the text is showing up in a quick reply.