bpo-45711: use exc_value instead of exc_type to determine if exc_info is valid. Add more assertions.

[iritkatriel]

This brings us closer to removing exc_type from exc_info.

Where exc_type is used, I made it use exc_value instead and added assertions that this is ok. Then when we remove exc_type it will just be a matter of removing the assertions.

https://bugs.python.org/issue45711

[bedevere-bot]

🤖 New build scheduled with the buildbot fleet by @iritkatriel for commit bc889bc 🤖

If you want to schedule another build, you need to add the ":hammer: test-with-buildbots" label again.

[gvanrossum]

I noticed that in some places you require type and value to be both "nullish" or both "not nullish" (where "nullish" means either NULL or Py_None), and in other places (in particular in _assert_exception_type_is_redundant()) they are required to both be the same nullishness (i.e., both NULL or both Py_None or neither of them nullish). Is there a reason for the difference?

[iritkatriel]

I noticed that in some places you require type and value to be both "nullish" or both "not nullish" (where "nullish" means either NULL or Py_None), and in other places (in particular in _assert_exception_type_is_redundant()) they are required to both be the same nullishness (i.e., both NULL or both Py_None or neither of them nullish). Is there a reason for the difference?

Yes. Most of these assertions are verifying that the change right next to them is valid. So if I replace if(exc_type is nullish) by if(exc_value is nullish) then I only care that their nullishness is the same. In the case of the triplets pushed to the stack, it used to be that no-exc was pushed as "NULL, NULL, None" (None for type, NULL for value and tb). I change that in this PR (ceval.c lines 5907-9 sorry - 4182-88) so now it's "NULL, None, None". There are places where "None at the top of the stack" means there is no exc_info. So when we will remove the type we need the value to be None and not NULL, otherwise we'll get segfaults.

[gvanrossum]

Okay, I'm through now. (Sorry for the premature send on the first batch.)

[gvanrossum]

While you're here could you change 'exc' and 'exc2' to 'type' and 'type2' (or 'typ' and 'typ2')? It drives me nuts that we mix "exc[exption]" and "typ[e]" for the same concept (this must date back to Python 1 when the value was not an exception instance and the exception could be any object).

[gvanrossum]

All good! (Alas, I'm still stuck trying to wrap my head around the except* implementation, so that one will probably have to wait until next week.)