STM32CubeIDE migration to 1.3.0 appears to break debugger

[etheory]

Hi there!

Each project I've migrated from the previous version of STM32CubeIDE to 1.3.0 can no longer debug properly.

When launching a debug session I now always get the following:

But if I create a new project from scratch everything is fine.

Is anyone else experiencing this behavior?

If so is there a setting I can set that will correct this?

Also my code no longer seems to get sent to the Nucleo board, unless it's from a project made with 1.3.0. Basically migrated projects seem broken.

Thanks.

[DEdwa.1]

I'm having exactly the same problems, and can't get the ST-LINK OpenOCD support to work at all when creating new projects either.

I think there is something very broken with v1.3.0. I even tried uninstalling it and installed a clean 1.3.0 incase it was the auto-update from 1.2.0 which caused the problem, but it hasn't improved things. As far as I've experienced v1.3.0 is not functional, and should probably be withdrawn.

I also tried creating a new workspace and creating a new default MX project from a reference board. This worked fine, but it failed to create a launch configuration - reporting internal errors for ST-LINK/GDB Server and missing files when trying to use the generally ar more robust OpenCD/GDB Server options.

Also when editing the debug configuration, it reports unsaved changes, clicking "save" doesn't indicate an error but doesn't save the configuration file ...

I've totally uninstalled STM32CUBEIDE and gone back to a fresh install of V1.2.0 and things are working as before.

This update has wasted many hours and I'm rather unhappy about it!

[KnarfB]

This happened to me after migration to 1.3.0 for existing projects / .elf files. Try cleaning the project and building from scratch.

It worked for me. Probably the underlying toolchain changed.

It might also be helpful to generate a new debug launch configuration.

[etheory]

Thanks I'll give that a go if it happens again, but luckily I've only just started with the ecosystem, so it was quite quick to just make a new project and copy over the relevant code.

For larger future projects I'll try your suggestion.