Repository navigation
Missing conversion from StringLiteral to unicode_type #6907
Description
Reporting a bug
- I have tried using the latest released version of Numba (most recent is
visible in the change log (https://github.com/numba/numba/blob/master/CHANGE_LOG). - I have included a self contained code sample to reproduce the problem.
i.e. it's possible to run as 'python bug.py'.
This:
from numba import typeof, types
from numba.core.typing import Context
x = types.literal('string literal')
print(x, type(x))
print("conversion:", x.can_convert_to(Context(), types.unicode_type))does this:
Literal[str](string literal) <class 'numba.core.types.misc.StringLiteral'>
conversion: None
I think it should be possible to convert a StringLiteral to a unicode_type but not the converse.
A can_convert_to() method needs adding to this class:
numba/numba/core/types/misc.py
Lines 40 to 41 in 3461bcb
xref: https://numba.discourse.group/t/how-to-pass-string-argument-to-numba-njit-python-function-from-inside-another-whilst-specifying-signature-type/663 which hits this bug.
Activity
gmarkall commentedon Apr 8, 2021
Would fixing this also fix #6897?
Add type conversion for StringLiteral to unicode_type and test.
Add type conversion for StringLiteral to unicode_type and test.
Add type conversion for StringLiteral to unicode_type and test.