GeezerVison

Hi @iritkatriel , could you take a quick look at the proposal and the implementation and see if this makes sense? I will add the tests and the documentation after your pre-approval.

Okay I defined a new Enum - PdbInvokeType for pdb to check whether to enable some commands internally. I also exported the enum because the debuggers based on pdb might need it.

Do you think code wise this is the way to go? I'll work on the docs and tests if the feature is good to go.

Hi @iritkatriel , I updated the docs and tests, also renamed the enum class. Could you review it? Thanks!

Hey @iritkatriel , I made some changes after the discussion on discord. Now the name of the argument is mode and it accepts a string 'cli' or 'inline', or a None.

A few questions - do you think we should add run-time checker for the value? If we do, will it affect devs that build their debugger based on pdb? Should we have a separate method to check it so it can be overwritten? We can also do a setting kind of thing.

The related question - should we type hint this? It's not common to type hint arguments in pdb.py - one of the reasons is of course the age of the file. However, last time I heard there's still a debate on whether we should type hint as much as possible in stdlibs.

I think one thing we need to factor in is that - this argument is basically exclusively used by debugger developers. The debugger users should never need to use this. So maybe we can cut some corners (not making it super safe to use).

For the documentation, for restart command, I did not link the availability of the command to the mode argument directly because that's not how normal user experiences the difference.

The doc change makes sense. I also added the whatsnew entry.