discuss multiple call to set_next_input

[Carreau]

See #323,

before big split, we were making sure each execution requests would not trigger multiple set_next_input, that was due to ? triggering multiple set_next_input in a for loop and annoying some users.

I can see how calling multiple set_next_input could make sens, but I don't think the protection code should be in the kernel.

[krassowski]

In JupyterLab we see that code like sum([1]).denominator? leads to unexpected an annoying behaviour (creating a next input instead of providing the help message), see jupyterlab/jupyterlab#12269. Is this issue related to it? Sorry I think I am missing some context, but I see a reference to ? syntax and I wonder if this also refers to%pinfo magic or is it a separate feature?