I'm no longer using the dialog. Instead, I created specific buttons for each operation but the old problem explained in my other post still remains.
For example, when I press a specific button, I would like the editor to break in a function 2 or 3 call later (in a different .py file than the one that started the processing). I find this restriction really working against developers trying to extract coherent code into specific modules (.py files) and reusing these modules as building blocks for more complex operations implemented in other .py files...
@stevebaer, you don't know how many times I would have gladly paid the performance price to be able to break even when the processing was not started form the editor! IMHO, this is a choice that should be put into the hands of the users/developers instead of automatically decided for them, even if used only during development. For example, in C++, you can decide to build your optimized code WITH the debug information still generated and usable...
BTW, I don't seem to have the Tracing Enabled option in the editor? See:
Here is my version info:
Version 5 SR14 32-bit
Well, that was not directly the point by I suspect that the same thing that was preventing the use of breakpoints in code used in release mode (not started from the editor) might also be stripping call stack line number information?
Thanks for your help, I'm still open to suggestions/workaround...