help
help > RE: Graph Theory Results Error
Dec 3, 2015 09:12 PM | Alfonso Nieto-Castanon - Boston University
RE: Graph Theory Results Error
Dear Pravesh,
Just a quick test while I try to figure out the root cause, the negative number in the [2 -262 2560 1339] line indicates that Matlab believes that your figure window is being cropped when maximized (it believes the window bottom side is positioned 262 pixels below the beginning of your screen), which is clearly not the case. Could you try, instead of maximizing the window using the corresponding icon/button on that window, manually resizing the window (e.g. drag&drop the window corner) until it reaches the full screen to see if the offset/scaling issue replicates under those conditions or not?
Also, would you happen to have something like a system-level taskbar placed at the bottom or top of your screen? (depending on the implementation unix taskbars some times interfere with the reported screen sizes to try to avoid other software from overlapping the taskbar)
Thanks
Alfonso
Originally posted by Pravesh Parekh:
Just a quick test while I try to figure out the root cause, the negative number in the [2 -262 2560 1339] line indicates that Matlab believes that your figure window is being cropped when maximized (it believes the window bottom side is positioned 262 pixels below the beginning of your screen), which is clearly not the case. Could you try, instead of maximizing the window using the corresponding icon/button on that window, manually resizing the window (e.g. drag&drop the window corner) until it reaches the full screen to see if the offset/scaling issue replicates under those conditions or not?
Also, would you happen to have something like a system-level taskbar placed at the bottom or top of your screen? (depending on the implementation unix taskbars some times interfere with the reported screen sizes to try to avoid other software from overlapping the taskbar)
Thanks
Alfonso
Originally posted by Pravesh Parekh:
Dear Dr.
Alfonso,
My apologies for not having been thorough in my testing. It seems that I forgot to restart MATLAB after putting the patch in (or else after playing around with the resolution). Restarting MATLAB has fixed the problem and now I can see the graph theory results (and hence, the patch works!)! Thank you so much!
Also, upon restarting MATLAB, groot now returns the current resolution (2560x1440). However, the scaling error is still present.
I am reposting the output from the lines:
973 604
2 -262 2560 1339
1 1 2560 1440
0.1756 0.5950 0.1350 0.3600
0 0 0 0
0 0 0 0
0.1350 0.9550 0.8650 0.0450
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0.9550 0.1350 0.0450
0.0100 0.0100 0.1150 0.1000
0 0.2700 0.1350 0.4800
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
Apologies again for the wrong information posted earlier. Do you have any suggestions on the scaling issue?
Thanks and Regards
Pravesh
Originally posted by Pravesh Parekh:
My apologies for not having been thorough in my testing. It seems that I forgot to restart MATLAB after putting the patch in (or else after playing around with the resolution). Restarting MATLAB has fixed the problem and now I can see the graph theory results (and hence, the patch works!)! Thank you so much!
Also, upon restarting MATLAB, groot now returns the current resolution (2560x1440). However, the scaling error is still present.
I am reposting the output from the lines:
973 604
2 -262 2560 1339
1 1 2560 1440
0.1756 0.5950 0.1350 0.3600
0 0 0 0
0 0 0 0
0.1350 0.9550 0.8650 0.0450
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0.9550 0.1350 0.0450
0.0100 0.0100 0.1150 0.1000
0 0.2700 0.1350 0.4800
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
Apologies again for the wrong information posted earlier. Do you have any suggestions on the scaling issue?
Thanks and Regards
Pravesh
Originally posted by Pravesh Parekh:
Dear Dr.
Alfonso,
Thank you for your quick reply.
a) The issue happens only during Graph theory results window (though occasionally the output gets spilled over to Second level window; since using release 15g, this spilling over has not been seen)
b) I have tried the patch but get the following error:
Error using matlab.graphics.primitive.Image/set
The UIContextMenu's parent figure is not the same as the parent figure of this object
Error in conn_displaynetwork (line 524)
uimenu(hc1,'Label','Display 3d view','callback',{@conn_displaynetwork,'display3d'});
Error in conn (line 5412)
conn_displaynetwork('init',CONN_x.Results.xX.nsources);
Error in conn_menumanager (line 119)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});
c) The original resolution of the monitor being used is 3840x2160. However, we have set it down to 2560x1440 because the font size gets too small on the max resolution. When I use groot on MATLAB, I am returned the 3840x2160 values. I tried changing to the maximum resolution and then starting and maximising Conn but the scaling issue is still there. Interestingly, the message boxes (for example, the loading box) appear right in the centre of the screen but Conn always starts in the lower right side of the screen.
The output:
Variable p1: 854 1241
Variable p2: 2 207 2560 1339
Variable p3: 1 1 3840 2160
The output on my MATLAB prompt:
854 1241
2 207 2560 1339
1 1 3840 2160
0 0 0 0
0.6081 0.7750 0.1350 0.1800
0 0 0 0
0.1350 0.9550 0.8650 0.0450
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0.9550 0.1350 0.0450
0.0100 0.0100 0.1150 0.1000
0 0.2700 0.1350 0.4800
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
(this was before the patch).
The following is the output after patch:
1255 1303
2 207 2560 1339
1 1 3840 2160
0.1756 0.5950 0.1350 0.3600
0.6081 0.7750 0.1350 0.1800
0.8244 0.7750 0.1350 0.1800
0.1350 0.9550 0.8650 0.0450
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0.6400 0.0990 0.3150
0 0.9550 0.1350 0.0450
0.0100 0.0100 0.1150 0.1000
0 0.2700 0.1350 0.4800
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0.0050 0.4400 0.1000 0.2400
0.0100 0.0600 0.1000 0.1000
0 0 0 0
0 0 0 0
On changing the screen resolution to 3840x2160, p2 changes value to: 2 -513 3840 2059
We do not have an extended display...I have checked the X11 config files but nothing seems out of ordinary...I will take a closer look and report back if something strikes me.
Thanks and Regards
Pravesh
Originally posted by Alfonso Nieto-Castanon:
Thank you for your quick reply.
a) The issue happens only during Graph theory results window (though occasionally the output gets spilled over to Second level window; since using release 15g, this spilling over has not been seen)
b) I have tried the patch but get the following error:
Error using matlab.graphics.primitive.Image/set
The UIContextMenu's parent figure is not the same as the parent figure of this object
Error in conn_displaynetwork (line 524)
uimenu(hc1,'Label','Display 3d view','callback',{@conn_displaynetwork,'display3d'});
Error in conn (line 5412)
conn_displaynetwork('init',CONN_x.Results.xX.nsources);
Error in conn_menumanager (line 119)
feval(CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});
c) The original resolution of the monitor being used is 3840x2160. However, we have set it down to 2560x1440 because the font size gets too small on the max resolution. When I use groot on MATLAB, I am returned the 3840x2160 values. I tried changing to the maximum resolution and then starting and maximising Conn but the scaling issue is still there. Interestingly, the message boxes (for example, the loading box) appear right in the centre of the screen but Conn always starts in the lower right side of the screen.
The output:
Variable p1: 854 1241
Variable p2: 2 207 2560 1339
Variable p3: 1 1 3840 2160
The output on my MATLAB prompt:
854 1241
2 207 2560 1339
1 1 3840 2160
0 0 0 0
0.6081 0.7750 0.1350 0.1800
0 0 0 0
0.1350 0.9550 0.8650 0.0450
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0.9550 0.1350 0.0450
0.0100 0.0100 0.1150 0.1000
0 0.2700 0.1350 0.4800
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
(this was before the patch).
The following is the output after patch:
1255 1303
2 207 2560 1339
1 1 3840 2160
0.1756 0.5950 0.1350 0.3600
0.6081 0.7750 0.1350 0.1800
0.8244 0.7750 0.1350 0.1800
0.1350 0.9550 0.8650 0.0450
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0.6400 0.0990 0.3150
0 0.9550 0.1350 0.0450
0.0100 0.0100 0.1150 0.1000
0 0.2700 0.1350 0.4800
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0.0050 0.4400 0.1000 0.2400
0.0100 0.0600 0.1000 0.1000
0 0 0 0
0 0 0 0
On changing the screen resolution to 3840x2160, p2 changes value to: 2 -513 3840 2059
We do not have an extended display...I have checked the X11 config files but nothing seems out of ordinary...I will take a closer look and report back if something strikes me.
Thanks and Regards
Pravesh
Originally posted by Alfonso Nieto-Castanon:
Hi
Pravesh,
Thanks for the detailed info, a couple of additional questions/thoughts:
a) does the same issue happen when displaying ROI-to-ROI results (ie. in the 'results explorer' button of the ROI-to-ROI second-level tab) or is it specific to the graph-theory results window?
b) could you please try the attached patch (this patch is for release 15g, copy the file to the conn distribution folder overwriting the file with the same name there) and see whether the UIContextMenu error message persists or no?
c) the "loss of spatial know-how" of the conn GUI when maximized looks similar to an old issue that appeared when connecting an external monitor or projector after CONN had already started (that issue has been fixed but it was related to a change in the screen resolution information that was only partially conveyed to Matlab), is there any reason why you think the screen size or resolution information might not match the actual size/resolution of your monitor (e.g. are you using some form of desktop extender, do you have multiple monitors, etc.)? Also if you could please run the following commands (after maximizing the CONN gui) and let me know their output that might help as well:
global CONN_MM;
p1=get(0,'pointerlocation');
p2=get(CONN_MM.gcf,'position');
p3=get(0,'screensize');
disp(p1);
disp(p2);
disp(p3);
disp(CONN_MM.POSITION);
Best
Alfonso
Originally posted by Pravesh Parekh:
Thanks for the detailed info, a couple of additional questions/thoughts:
a) does the same issue happen when displaying ROI-to-ROI results (ie. in the 'results explorer' button of the ROI-to-ROI second-level tab) or is it specific to the graph-theory results window?
b) could you please try the attached patch (this patch is for release 15g, copy the file to the conn distribution folder overwriting the file with the same name there) and see whether the UIContextMenu error message persists or no?
c) the "loss of spatial know-how" of the conn GUI when maximized looks similar to an old issue that appeared when connecting an external monitor or projector after CONN had already started (that issue has been fixed but it was related to a change in the screen resolution information that was only partially conveyed to Matlab), is there any reason why you think the screen size or resolution information might not match the actual size/resolution of your monitor (e.g. are you using some form of desktop extender, do you have multiple monitors, etc.)? Also if you could please run the following commands (after maximizing the CONN gui) and let me know their output that might help as well:
global CONN_MM;
p1=get(0,'pointerlocation');
p2=get(CONN_MM.gcf,'position');
p3=get(0,'screensize');
disp(p1);
disp(p2);
disp(p3);
disp(CONN_MM.POSITION);
Best
Alfonso
Originally posted by Pravesh Parekh:
Hi,
Sorry for raking up old threads but I have been running into the same error as Stephen and Charles.
What I have noticed is as follows:
- Issue happens on (Red Hat) Linux (I am able to open the same project on Windows and see the graph theory results)
- Unlike Stephen, I get consistent errors irrespective of the contrast selected
- I have noticed that even when loading Results Explorer window, Conn tends to "draw" the results in the second-level tab rather than Results explorer window (eventually moves to results window on updating)
- When the window is maximized, Conn sort of loses its spatial know-how. Clicks get way off the mark (i.e. you have to click in apparent empty spaces to actually select something way up; screenshot attached; have added a red circle to the original pointer location)
- Error is present across different projects (and conditions/number of subjects/ROIs)
- Error is persistent irrespective of the number of windows open (i.e results explorer, seed-to-voxel results...)
- Error, mis-placed drawing results, and scaling issues persist across versions 15e and 15g
- In the graph theory results, conn is unable to even remove the axes (the tick marks are visible with the error message)
- I am assuming that the patch was already incorporated in 15g (hence have not tried it explicitly)
- Again, unlike Stephen, the system is not configured in a cluster environment (though its connected to other systems over the network)
- The system is running with Nvidia Quadro K620 GPU (and the drivers are up-to date; issue persisted across driver update)
- MATLAB version is 2015a with SPM12b
I am wondering if this has something to do with the Nvidia server-x configuration? We have experienced display issues in the past because of a server-x configuration...or perhaps Conn is reading something that is differently returned by Windows/Linux?
Thoughts/suggestions on how to proceed?
The error reads as follows:
Regards
Pravesh Parekh
Sorry for raking up old threads but I have been running into the same error as Stephen and Charles.
What I have noticed is as follows:
- Issue happens on (Red Hat) Linux (I am able to open the same project on Windows and see the graph theory results)
- Unlike Stephen, I get consistent errors irrespective of the contrast selected
- I have noticed that even when loading Results Explorer window, Conn tends to "draw" the results in the second-level tab rather than Results explorer window (eventually moves to results window on updating)
- When the window is maximized, Conn sort of loses its spatial know-how. Clicks get way off the mark (i.e. you have to click in apparent empty spaces to actually select something way up; screenshot attached; have added a red circle to the original pointer location)
- Error is present across different projects (and conditions/number of subjects/ROIs)
- Error is persistent irrespective of the number of windows open (i.e results explorer, seed-to-voxel results...)
- Error, mis-placed drawing results, and scaling issues persist across versions 15e and 15g
- In the graph theory results, conn is unable to even remove the axes (the tick marks are visible with the error message)
- I am assuming that the patch was already incorporated in 15g (hence have not tried it explicitly)
- Again, unlike Stephen, the system is not configured in a cluster environment (though its connected to other systems over the network)
- The system is running with Nvidia Quadro K620 GPU (and the drivers are up-to date; issue persisted across driver update)
- MATLAB version is 2015a with SPM12b
I am wondering if this has something to do with the Nvidia server-x configuration? We have experienced display issues in the past because of a server-x configuration...or perhaps Conn is reading something that is differently returned by Windows/Linux?
Thoughts/suggestions on how to proceed?
The error reads as follows:
Error using
matlab.graphics.primitive.Image/set
THe UIContextMenu's parent figure is not the same as the parent figure of this object
Error in conn_displaynetwork(line 524)
set(data.refaxes,'uicontextmenu',hc1);
Error in conn_menumanager (line 119)
feval((CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});
THe UIContextMenu's parent figure is not the same as the parent figure of this object
Error in conn_displaynetwork(line 524)
set(data.refaxes,'uicontextmenu',hc1);
Error in conn_menumanager (line 119)
feval((CONN_MM.MENU{n0}.callback{n1}{1},CONN_MM.MENU{n0}.callback{n1}{2:end});
Regards
Pravesh Parekh
Threaded View
| Title | Author | Date |
|---|---|---|
| Charles Malpas | Jan 20, 2015 | |
| Stephen Bailey | Sep 23, 2015 | |
| Stephen Bailey | Sep 23, 2015 | |
| Alfonso Nieto-Castanon | Sep 25, 2015 | |
| Stephen Bailey | Oct 8, 2015 | |
| Pravesh Parekh | Dec 3, 2015 | |
| Alfonso Nieto-Castanon | Dec 3, 2015 | |
| Pravesh Parekh | Dec 3, 2015 | |
| Pravesh Parekh | Dec 3, 2015 | |
| Alfonso Nieto-Castanon | Dec 3, 2015 | |
| Alfonso Nieto-Castanon | Dec 3, 2015 | |
| Pravesh Parekh | Dec 3, 2015 | |