Bad display and Crash when viewing a diagram in plotting perspective

[ceraolo]

I often use diagram display in plotting perspective. This feature has some internal bug, since when I use it a crash often occurs after some usage.

Now I've found a case that probably is related. Consider the model

PowerGrids.Examples.IEEE14bus.IEEE14busLine1to5Opening.

When opened, it shows very badly in plotting perspective: only the BoleanSOurce block is shown (see enclosed impage.png).

After this very bad display, OMEdit typically crashes very soon: I've just sent a report a few minutes ago.

Done using v1.16-0-dev.03

[ceraolo]

@adeas can you see if you can replicate this (I mean you see something similar to Image.png)?
In case you do, you'll probably be able to find a serious crash-causing bug in OM.
And that would enhance OMEdit usability significantly.

[casella]

I just tried this out. I confirm that the diagram in the plotting perspective is broken, as it only shows the signal source and a piece of connection line. Though it is a bit different in my case, maybe @ceraolo is still on Windows 7?

I'm not getting any crashes myself, at least not after a few minutes.

[ceraolo]

Replying to casella:

I just tried this out. I confirm that the diagram in the plotting perspective is broken, as it only shows the signal source and a piece of connection line. Though it is a bit different in my case, maybe @ceraolo is still on Windows 7?

Win 10 build 1909

[adeas31]

I get the same behavior as reported by @casella.

[adeas31]

The display issue is fixed in e15e9e2/OpenModelica.

[ceraolo]

I've just tried.
Got the diagram, but also an immediate crash.
I sent the crash report. It should be good for the analysis, since the actions I made were just

• open OMEdit
• load PowerGrids
• open PowerGrids.Examples.IEEE14bus.IEEE14busLine1to5Opening
• go in the plotting perspective and show the diagram
• close the diagram (in plotting perspective) -> CRASH

[ceraolo]

Since there is more than just a suspect that there is some relationship between plot perspective display and crash, I think it is better to keep this ticket open.
@adeas, I'm looking forward to seeing what you can in infer from today's crash report.

[casella]

I tried this myself. I confirm that I can now see the diagram also in the plotting perspective. I opened and closed it multiple times, no problem.

I'm using Windows Version 10.0.18363 build 18363.

@ceraolo, build 1909 looks pretty old. Maybe you should run a few system updates?

[adeas31]

I can reproduce the crash now. It is crashing on the close.
Should be fixed with ​PR#6666.

[adeas31]

Fixed in 81768a1/OpenModelica.

[ceraolo]

Replying to casella:

I tried this myself. I confirm that I can now see the diagram also in the plotting perspective. I opened and closed it multiple times, no problem.

I'm using Windows Version 10.0.18363 build 18363.

@ceraolo, build 1909 looks pretty old. Maybe you should run a few system updates?

I'm as updated as I can on all of my PCs.
Indeed, I used "build", which is not the correct name. I should have used "Version" instead.
There are two versions currently on updated win10 PCs. On my desktop, I have the most recent one, i.e #2004.
My laptop has instead the previous one, i.e. 1909. These days Microsoft is releasing 2004 progressively, depending on the hardware installed on a PC. For My latop's hardware, it says win10 is not ready yet to update to 2004.

[ceraolo]

Replying to adeas31:

I can reproduce the crash now. It is crashing on the close.
Should be fixed with ​PR#6666.

Great!
I hoped that a crash after so few steps could tell the tale.
Finding crash-causing issues is always a great satisfaction!
I'll check tomorrow.

[casella]

Very good. It remains a mistery to me why I couldn't reproduce that, but all's well that ends well :)

@ceraolo, please close the ticket yourself if the next nightly no longer has the issue.

[ceraolo]

Replying to casella:

Very good. It remains a mistery to me why I couldn't reproduce that

Crashes are often erratic.
In some cases, for instance, bugs cause to write to an unallocated area. In this case, the result depends on what is there (I.e. what we overwrite), and this can be different from a PC to another, between runs on the same PC, or even in the same run for the same user action depending on what has been done before.
The ticket is already closed.