Fabio Zadrozny

Ok, I've just fixed the tests and added back methods which had been removed from the public API. Also, lazy imports are added very judiciously in the PR, so, I...

One thing that just occurred to me is that this is probably not ideal for the language server as it needs the source/line/col for the variable. As a note, right...

Seems good to me -- `+1` on delaying on the chance of having that information later on ;)

Regarding the monkey-patching, I'm a bit wary of it as it's bound to break as the debugger moves forward (if this functionality is really needed, this should probably be in...

Humm, if you know the points when you want to start/stop tracing and want absolute control, maybe using `get_global_debugger().enable_tracing()` / `get_global_debugger().disable_tracing()` could be a better approach...

I couldn't find any reference to `__ipython_bottom__` (apart from the usage where it's checked in the debugger). Where would that be defined? p.s.: the enable/disable tracing would still leave items...

> * Would we need to be careful to reapply the same tracers? I.e. would the trace state need to be stored between calls to `disable` and the next `enable`?...

After thinking a bit more about it I think that your original approach of having a root frame from which the debugging happens (relying on `__tracebackhide__ == "__ipython_bottom__"`) may really...

@vidartf I've added some thoughts on the plans to tackle this in the debugger (see: https://github.com/microsoft/debugpy/issues/869#issuecomment-1119864916). After I finish tackling that issue, I'll give you a heads up so that...

Hi, thanks for the suggestions, still, I haven't been able to make it much better... I can make it a little less wider by making the spacing smaller (but that's...