Resizing the APP window makes it freeze and automatically closes it.

[wtfrules]

Hi,

Using the latest version, manually resizing the APP window makes it freeze and automatically closes it.
No matter if it´s connected to a multimeter or no.
I´m using Windows 11.-

Attached is a video showing the problem.

Thanks

rustymeter.mp4

[markusdd]

That is extremely interesting. When no multimeter is connected that kind of rules out the serial background process.

Considering that, it would seem either a graphics driver issue or something with egui that does not play together with your system.

I just tried and cannot reproduce it on my Win11 installation. Also not on Linux.

Do you feel comfortable building it yourself in debug mode and running it so we can see where the error occurs?

[wtfrules]

Hi @markusdd , yes, this is odd, I was using v0.4.0 and this issue wasn´t there lol.
In v0.4.4 this issue persists with or without multimeter.
I´ll try to build it per your suggestion and provide feedback.

Thanks

[wtfrules]

OK @markusdd , after building it in my computer, I´ve the same issue resizing the window of the app...

Rusty.Meter.mp4

[markusdd]

Ok cool. When you build with debug and run it via the terminal, do you get a crash dump/ stack trace?

This is what we need to see whether it is in my code or something in egui or graphics

[wtfrules]

mmm, seems I just build it per instructions from project main page, so how do I build it with and run it with debug features?

Thanks

[markusdd]

If you remove --release you should get a debug build. When starting there will be a small orange indicator of that in the main window saying Debug Build.

If you then launch the program using the terminal and not from explorer, you should get errors printed when it crashes

[wtfrules]

well, still no luck, with the debug version recompiled in my system, still crashes and no debug data....lol

Rusty.Meter.mp4

[markusdd]

that is...wow. xD

can you run echo %errorlevel% afterwards to see if the exitcode is not 0?

If absolute nothing is printed maybe the console out stream is not attached.

Can you retry after removing line 3 from src/main.rs? Maybe we are suppressing too much console printing on windows (this line exists so you do not get a ghost console window each time)

Also please try running it with the Backtrace variable enabled: RUST_BACKTRACE=1 ./rusty_meter

[markusdd]

Asked over at egui where there is a ticket that sounds similar: emilk/egui#8043

[wtfrules]

that is...wow. xD

can you run echo %errorlevel% afterwards to see if the exitcode is not 0?

If absolute nothing is printed maybe the console out stream is not attached.

Can you retry after removing line 3 from src/main.rs? Maybe we are suppressing too much console printing on windows (this line exists so you do not get a ghost console window each time)

Also please try running it with the Backtrace variable enabled: RUST_BACKTRACE=1 ./rusty_meter

Hi, sorry for the delay.

recompiled deleting line 3 of src/main.rs

now I got this:

C:\Users\WTF Rules!\rusty_meter\target\debug>echo %errorlevel%
0

C:\Users\WTF Rules!\rusty_meter\target\debug>rusty_meter.exe

C:\Users\WTF Rules!\rusty_meter\target\debug>echo %errorlevel%
-1073740771

but still no debug info, and the command :RUST_BACKTRACE=1 ./rusty_meter

gives error about not finding that as an executable or so.

Thanks

[markusdd]

Ah. The command probably fails because this is bash Syntax to run a command with a variable.

Do you have git-bash on your machine by any chance, you should if you have git? With that it should work this way.

If you have VS Code you can also launch a bash terminal from within it.

[wtfrules]

I dont have them, just not a programmer lol

[markusdd]

can you try the original syntax with RUST_BACKTRACE=1 prepended in powershell instead of cmd?

In cmd it probably has to be done like so:

set RUST_BACKTRACE=1
rusty_meter.exe

[wtfrules]

Okay, I tried my best, lol