Re: [PATCH 08/19] qapi/schema: add static typing and assertions to lookup_type()

[John Snow]

On Wed, Nov 22, 2023 at 7:09 AM Markus Armbruster <armbru@redhat.com> wrote:
>
> John Snow <jsnow@redhat.com> writes:
>
> > On Tue, Nov 21, 2023, 9:21 AM Markus Armbruster <armbru@redhat.com> wrote:
> >
> >> John Snow <jsnow@redhat.com> writes:
> >>
> >> > This function is a bit hard to type as-is; mypy needs some assertions to
> >> > assist with the type narrowing.
> >> >
> >> > Signed-off-by: John Snow <jsnow@redhat.com>
> >> > ---
> >> >  scripts/qapi/schema.py | 8 ++++++--
> >> >  1 file changed, 6 insertions(+), 2 deletions(-)
> >> >
> >> > diff --git a/scripts/qapi/schema.py b/scripts/qapi/schema.py
> >> > index a1094283828..3308f334872 100644
> >> > --- a/scripts/qapi/schema.py
> >> > +++ b/scripts/qapi/schema.py
> >> > @@ -968,8 +968,12 @@ def lookup_entity(self, name, typ=None):
> >> >              return None
> >> >          return ent
> >> >
> >> > -    def lookup_type(self, name):
> >> > -        return self.lookup_entity(name, QAPISchemaType)
> >> > +    def lookup_type(self, name: str) -> Optional[QAPISchemaType]:
> >>
> >> Any particular reason not to delay the type hints until PATCH 16?
> >>
> >
> > I forget. In some cases I did things a little differently so that the type
> > checking would pass for each patch in the series, which sometimes required
> > some concessions.
> >
> > Is this one of those cases? Uh, I forget.
> >
> > If it isn't, its almost certainly the case that I just figured I'd type
> > this one function in one place instead of splitting it apart into two
> > patches.
> >
> > I can try to shift the typing later and see what happens if you prefer it
> > that way.
>
> Well, you structured the series as "minor code changes to prepare for
> type hints", followed by "add type hints".  So that's what I expect.
> When patches deviate from what I expect, I go "am I missing something?"
>
> Adding type hints along the way could work, too.  But let's try to stick
> to the plan, and add them all in PATCH 16.
>
> >> > +        typ = self.lookup_entity(name, QAPISchemaType)
> >> > +        if typ is None:
> >> > +            return None
> >> > +        assert isinstance(typ, QAPISchemaType)
> >> > +        return typ
> >>
> >> Would
> >>
> >>            typ = self.lookup_entity(name, QAPISchemaType)
> >>            assert isinstance(typ, Optional[QAPISchemaType])
> >>            return typ
> >>
> >> work?
> >>
> >
> > I don't *think* so, Optional isn't a runtime construct.
>
> Let me try...
>
>     $ python
>     Python 3.11.5 (main, Aug 28 2023, 00:00:00) [GCC 12.3.1 20230508 (Red Hat 
> 12.3.1-1)] on linux
>     Type "help", "copyright", "credits" or "license" for more information.
>     >>> from typing import Optional
>     >>> x=None
>     >>> isinstance(x, Optional[str])
>     True
>     >>>

Huh. I ... huh!

Well, this apparently only works in Python 3.10!+

TypeError: Subscripted generics cannot be used with class and instance checks

>
> >                                                         We can combine it
> > into "assert x is None or isinstance(x, foo)" though - I believe that's
> > used elsewhere in the qapi generator.
>
> >> >
> >> >      def resolve_type(self, name, info, what):
> >> >          typ = self.lookup_type(name)
> >>
> >>
>