You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
error: locale.bind_textdomain_codeset is not present at runtime
Stub: at line 147
def (builtins.str, Union[builtins.str, None]) -> Union[builtins.str, None]
Runtime:
MISSING
error: locale.bindtextdomain is not present at runtime
Stub: at line 146
def (builtins.str, Union[Union[builtins.str, os.PathLike[builtins.str]], None]) -> builtins.str
Runtime:
MISSING
error: locale.dcgettext is not present at runtime
Stub: at line 144
def (Union[builtins.str, None], builtins.str, builtins.int) -> builtins.str
Runtime:
MISSING
error: locale.dgettext is not present at runtime
Stub: at line 143
def (Union[builtins.str, None], builtins.str) -> builtins.str
Runtime:
MISSING
error: locale.gettext is not present at runtime
Stub: at line 142
def (builtins.str) -> builtins.str
Runtime:
MISSING
error: locale.textdomain is not present at runtime
Stub: at line 145
def (Union[builtins.str, None]) -> builtins.str
Runtime:
MISSING
error: socket.MsgFlag.MSG_NOSIGNAL is not present in stub
Stub:
MISSING
Runtime:
MsgFlag.MSG_NOSIGNAL
#8096 just got a 3.10.4 run and passed (after failing #8096 (comment)). Maybe Github reverted something? Or a gradual rollout? I didn't see any relevant 3.10.5 changes from a skim
Okay, built 3.10.5 locally on my Mac and the locale ones do indeed seem like false positives. So seems like a problem with the Github build?
The socket one seems like a true positive (but it exists for me on 3.10.4 as well).
If we get nondeterministic failures, note that there's a nice stubtest allowlist trick we can use with an optional group to avoid unused allowlist entry complaints: https://github.com/python/mypy/blob/eb1b1e007e7fd1e976e6dd0f49a71b662a30a2d6/mypy/stubtest.py#L1459
Activity
hauntsaninja commentedon Jun 19, 2022
#8096 just got a 3.10.4 run and passed (after failing #8096 (comment)). Maybe Github reverted something? Or a gradual rollout? I didn't see any relevant 3.10.5 changes from a skim
hauntsaninja commentedon Jun 19, 2022
Okay, built 3.10.5 locally on my Mac and the locale ones do indeed seem like false positives. So seems like a problem with the Github build?
The socket one seems like a true positive (but it exists for me on 3.10.4 as well).
If we get nondeterministic failures, note that there's a nice stubtest allowlist trick we can use with an optional group to avoid unused allowlist entry complaints:
https://github.com/python/mypy/blob/eb1b1e007e7fd1e976e6dd0f49a71b662a30a2d6/mypy/stubtest.py#L1459
socket: consistency between constants and enums
Silence mysterious 3.10.5 stubtest failure
socket: consistency between constants and enums (#8099)
Silence mysterious 3.10.5 stubtest failure (#8102)