python / python/mypy

SupportsRound[int] hinted functions work with floats if round is called with no additional args

Offen
#8,213 1 Kommentar 0 Reaktionen 0 zugewiesene Personen Auf GitHub ansehen

Dieses Issue hat noch niemand übernommen.

bug
Vorherrschende Sprache
Python
Sterne
20.6k
Forks
3.3k
PR-Merge-Kennzahlen
PR-Kennzahlen ausstehend

Beschreibung

Note: if you are reporting a wrong signature of a function or a class in
the standard library, then the typeshed tracker is better suited
for this report: https://github.com/python/typeshed/issues

this genuinely might be a typeshed issue i'm not 100% sure. To be sure some similar issues about __round__ have cropped up both here and in typeshed:

#6211
python/typeshed#2638

If maintainers think this is solely a typeshed issue and would like me to repost over there please just let me know. If it is a typeshed issue and is a regression ( which it may be) then perhaps I could help by adding a test somewhere?

Are you reporting a bug, or opening a feature request?

possible bug

Example:
import typing

def round_it(x: typing.SupportsRound[int]) -> int:
    return round(x)

round_it(3.2)
What is the actual behavior/output?

though round is called with no ndigits arg in the example above we still get the following error when running mypy:

scratch.py:11: error: Argument 1 to "round_it" has incompatible type "float"; expected "SupportsRound[int]"
scratch.py:11: note: Following member(s) of "float" have conflicts:
scratch.py:11: note:     Expected:
scratch.py:11: note:         @overload
scratch.py:11: note:         def __round__(self) -> int
scratch.py:11: note:         @overload
scratch.py:11: note:         def __round__(self, ndigits: int) -> int
scratch.py:11: note:     Got:
scratch.py:11: note:         @overload
scratch.py:11: note:         def __round__(self, ndigits: None = ...) -> int
scratch.py:11: note:         @overload
scratch.py:11: note:         def __round__(self, ndigits: int) -> float

note that even if I modify this bit of typeshed to include :

@overload
def __round__(self, ndigits: None = ...) -> int: ...

I still have the same issue ( just with <1 more overload not shown>, etc)

What is the behavior/output you expect?

since SupportsRound is parameterised by return type and I have called the round method with no extra args I would expect that the example provided above would pass.

Specifically :
given both float and SupportsRound[int] more or less share a signature for the def __round__(self) and round is called w/ no args. I would expect mypy to be able to interpret that we're using def __round__(self, ndigits: None = ...) -> int and not care about the conflicting overload since it's not called.

What are the versions of mypy and Python you are using?

python: 3.8.0
mypy: 0.761 (and master)

What are the mypy flags you are using?

I have tried running with each of the following flag sets :

  • no flags
  • --strict
  • --no-implicit-optional
  • --no-strict-optional
  • --strict-optional

Beitragsleitfaden

Beitragsleitfaden öffnen

Erste Schritte

  1. Lies das ganze Issue und danach den Beitragsleitfaden des Projekts.
  2. Schreib ins Issue, dass du es übernimmst — das erspart doppelte Arbeit.
  3. Forke das Repository und arbeite in einem Branch.
  4. Öffne einen Pull Request, der die Issue-Nummer nennt.

Rechercherichtung

Reproduziere das bereitgestellte scratch.py-Beispiel mit mypy 0.761 oder master und untersuche anschließend die verknüpfte typeshed-Definition von round in stdlib/2and3/builtins.pyi. Bestimme, ob die Abweichung in mypy oder typeshed behoben werden sollte, und vergleiche das Verhalten mit den verwandten Issues. Erledigt bedeutet, dass die Zuständigkeit bestätigt ist und der erwartete round()-Aufruf durch einen geeigneten Test oder eine Übergabe abgedeckt ist.

Vom Indexierungsmodell aus dem Issue-Text verfasst.

Bewertung

Tech-Stack
python
Bereich
devtools
Issue-Typ
Bug
Schwierigkeit
4/5
Geschätzter Aufwand
3-5 Tage
Aktivitätsstatus
Veraltet
Klarheit
Größtenteils klar
Anfängerfreundlichkeit
35/100

Neue Issues direkt in Ihr Postfach

Eine kurze Übersicht über anfängerfreundliche GitHub-Issues.