I face the same problem as Chris. 

Suddenly my GDBStub project on my Arduino Nano board starts to pop up with the same message as described here. It used to work before so actually I am a bit confused here.

Before I have never faced miDebuggerPath and still I haven't any knowledge on this. From the post here I have searched for and found the file searched for and found the file : debugger_launch.json on my disc. But as my project is set up to GDB Stub (see the picture below) this file shouldn't be of any interest in my case I suppose.
 
If I set the Hardware to Manual/Custom in my project I got antoher error : Unable to start debugging. Launch options string provided by the project is invalid. The string is null or empty. 

If you could come up with a 2 line explanation of the miDebbuggerPath issue, it would be much appreciated. Perhaps even a suggestion on what setup in a system could possible have damaged something regarding miDebuggerPath.

Ok. I have done the steps you asked for and have attached the text file to this message. 

Having a quick look at this file myself, it seems to me that the MiDebuggerpath is not correct. But actually I can't remember, if I have set this us at some point working with this issue.

After reviewing this is likely an issue with the GDB exe not being included in the board package.

Can you try updating Visual Micro to the latest version, and retry the process as before?
(attach the log again if it fails so we can see any other issues)

Thanks for the detail.

Can you confirm all the additions for the GDBStub library are still there (header, debug_init(). breakpoint()?

Is there any further shown in the Output > Debug Window when this error happens?

Thanks for confirming, and the detail.

Do you get the same result whether you are using both "Debug > Start Debugging" and "Debug > Attach to Process" ?

Thanks for confirming and the additional information.

The same sketch should also run on an Uno if you have one?

A re-install may help, however it seems all files are present and in the right place/versions from the logs so far. Trying a different board may show it is all working and as you say there is another issue involved.

Hi again
I have now tried one of my ArMega168 boards. Unfortunately with the same result as I got when using my Arduino Nano board. 
A point of interest though. I found out that for both boards (AtMega168 and Arduino Nano) the SW seems to reset when the statement  : breakpoint(); in top of the loop function is reached. 
When removing this instruction for both projetcs, the SW runs as expected (the Internal LED is blinking).

Ok regarding the AtMega168 devices. We actually plan to fade these devices out in the nearest future, so that wouldn't be any problem for us. 
I have just tried the Serial debugging mode now just for the purpose. And of course compared to not having any debugging possibilities at all, this is a step in the right direction. 
Still though we aim for the GDB debugging possibilities. This is due to the facts that :

1) It is possible to single step through a program
2) Value of Variables can be watched
3) The one breakpoint you can set (tested on a NodeMCU 8266 device, I don't know if more breakpoints can be set on other devices) is moveable, without needing to download the code again.

So the "only" difference, as I see it with my knowledge at this moment, when comparing GDB Debugging with JTag debugging is the options being able to set more than one breakpoint when using a JTag debugger.   

Actually we plan to replace all out current devices with ES32 devices and using JTag debuggers with these devices. I don't hve any JTAg debugger for the one ESP32 device I have right now, so I don't have any possibilities to test debugging capabilities at my own PC. But as far as I can see it from Youtube videos and your own videos, it should be quite easy to get this setup up running. And is it also correct understood, that it is not possbile to use GDB Debugging, when working with ESO32 devices. Here you should have a JTag debugger at hand, if you want to use HW debugging ?

Though we aim for the ESP32 devices, still I will very much like to have the GDB Debugging for the Arduino Nano working. I actually need a working demo for this device to convince my colleagues, that Visual Micro is the tool we should choose. 

I had to switch back to my Arduino Uno board due to the fact, that I have a problem uploading the new  optiboot_atmega328,hex file from the Arduino IDE packages to my Arduino Nano board.
Doing this and still having the Device in visual Micro selected as : Atmega328P (Arduino Nano) I am able to upload my small code example to the board in Visual Micro and I am able to get the GDB Debugger up running to a certain point. When the breakpint statement in top of my loop functions is excuted, I receive a Trace/breakpint trap error. I can run/step from here, but it seems like a variable im at testing on in the watch window is only updated every second time.   
The problem is the same if I chosse Arduino Uno as device.

I have attached the output from the Debug window to this post.   

Yes it is still the code updating the variable on two consecutive lines. 
But I "made a small mistake" here concerning the variable NumberOfTimesInLoop. I discovered this having a closer look at my small piece of code. If I declare the variable as being volatile, I can see the variabel update in the Watch window on each statement updating the variable.

So the only problem left is the one with the Trace/Breakpoint error. If this can be solved, I will be a happy man