ports: Layout and content of the information shown by repr(object).

[robert-hh]

At the moment, the result of repr(object) when object is one created by the machine classes (and others) like a Pin object or UART object is somewhat inconsistent between the ports. And I contributed my share to the chaos. This inconsistency make it difficult to retrieve information if needed. So I suggest to harmonize that, knowing that such a change is both breaking and quite a bit of work. Not complicated, but a lot.
First thing would be to agree on the format. my suggestion would be the style and content used to instantiate the object in that port, with a minimum set of common information, and optional port specific information. So for a pin the result would e.g. be: pin-spec, mode=Pin.mode, pull=Pin.mode like 2, mode=Pin.OUT, pull=Pin.OFF. The pin-spec would have the form and content such that is can be used again to designate a pin, with a preference for a single integer. UART e.g. would be:
id, baudrate, tx=Pin(pin-spec), rx=Pin(pin-spec), bits=n, parity=UART.mode, stop=n, txbuf=n, rxbuf=n, ....,
PWM could be Pin(pin-spec), freq=nnn, duty_u16=nnn, invert=0|1, ..... No additional brackets of specifiers for a device type.
Using that, it is easier to parse such an information string and extract the required information.

[rkompass]

Using that, it is easier to parse such an information string and extract the required information.

I seem to have difficulties to parse such a string. I was trying to get a args, kwargs pair.
Say r = repr(obj)
Then args = tuple((int(a) for a in r.split(', ') if '=' not in a)) gives me args.
kwargs=eval(f'dict({','.join((a for a in s2.split(', ') if '=' in a))})') works but does not looks elegant.