I’m trying to write a simple bash script that opens up GQRX, sends it some TCP commands, then closes it down.
Unfortunately, I’ve found that when I close the program like this, the next time it opens, it will pop up a window saying “crash detected” and ask me to review the configuration file. This prevents the app from loading unless someone is present to click the dialog box.
This error only seems to happen when I try to close the program using the bash script. Closing it by just clicking the X doesn’t cause this problem next time it’s launched.
I think I’m closing the app too aggressively which terminates it before it can wrap up its affairs, and it interprets this as a crash. What’s the best way to close the app to keep this from happening?
I’ve tried:
- pkill -3 gqrx
- pkill -13 gqrx
But the problem persists. Is there an even softer way to close an application?
SIGTERM?
pkill -15 gqrxWould have been my suggestion as well.
No dice.
Removed by mod
could also try SIGTERM (15)
There’s a setting called ‘crashed’ in the config that is set to false when the main window destructor is called https://github.com/gqrx-sdr/gqrx/blob/master/src/applications/gqrx/mainwindow.cpp
You may be able to copy a new config file to reset this variable after killing the app
Good find! This might be a good backup plan.
Or sed the existing config file at the end of your script.
My first thought would be to use xdotool. Does that work for you?
This, or
wmctrlthat has some overlap. OR if this is Wayland (and not Xorg), whatever alternatives for those the window manager provides.Anyone know what those are?
Yeah agreed. All the people in here talking about signals… you’re not, like, wrong, but you missed the nature of the question i.e. “hey it seems like this GUI application doesn’t deal well with Unix signals, so what else can I do”
The program will detect that it was closed via the SSD / CSD decorations (the three little buttons, one with the x) and handle the killing itself.
Maybe you can invoke that close command through the app?
Closing from the X icon on the app window lets the program write out any info it needs to before shutdown. Killing the process from the terminal just ends the process totally
That is not true. If you use SIGKILL (9) it ends the process totally. Other signals don’t do that. And per default the cli commands don’t send SIGKILL, but SIGTERM. It nicely asks the program to terminate and the program is supposed to write down it’s data and do what it needs to do when signaled like that. You’d need to attach a ‘-9’ to kill it without warning. (Even if the command is called ‘kill’.)
A lot of GUI apps unfortunately don’t catch SIGTERM, so it doesn’t make much of a difference. A lot of them put their “exited normally” into the “the window was closed” handler.
For GQRX that might even be intentional, it’s very unstable at times, and you do kill it pretty often. So they probably assume if it got SIGTERM, it was frozen and the WM offered the user to kill it and clicked yes.
Sure,
if GUI apps don’t handle it they just keep running. It’s just asking them to terminate.[Edit: I’ll try that](We could just have a look at the code of GQRX in specific or just try it.)
The default behaviour for SIGTERM (and most other signals) when unhandled is to terminate the process, so you might as well be sending SIGKILL. It won’t keep running unless specifically coded to catch the signal and do nothing with it.
then it sounds like what is being killed has another child or parallel process that isn’t getting killed by the command used. Or Sigterm 15 is needed for graceful shutdown of process
It’s more than an “ask” - if it’s not handled then the application will still shutdown. SIGTERM is “trapable” by the application though and should be used to safely terminate.
They’ll also want to try SIGINT as that’s what ctrl+c sends. The application may handle that one instead.
Only if you -9 it
