Fix bool to int casting issue in Python
#1654
Conversation
|
CUDA Quantum Docs Bot: A preview of the documentation can be found here. |
| if F64Type.isinstance(left.type) and IntegerType.isinstance( | ||
| comparator.type): | ||
| left = arith.FPToSIOp(comparator.type, left).result |
There was a problem hiding this comment.
I'm not following why the LHS, if it's a floating point value, is integer truncated here.
There was a problem hiding this comment.
This is existing code from the isinstance(op, ast.Eq) below, which, in principle, should also be handled in the ast.NotEq case here.
I don't know the context of it so didn't change it or migrate it to cc.CastOp yet.
@khalatepradnya Do you know the intent of this float -> int conversion in equal check?
There was a problem hiding this comment.
nope, sorry. Tagging @amccaskey for details.
I see that the comparator is of integer type and that the integer comparator is being used (arith.cmpi) on line#2740
There was a problem hiding this comment.
I'm wondering if this isn't a bug. Consider this python code
>>> a=1
>>> b=1.2
>>> c= b==a
>>> print(c)
False
|
CUDA Quantum Docs Bot: A preview of the documentation can be found here. |
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>
|
CUDA Quantum Docs Bot: A preview of the documentation can be found here. |
bettinaheim
changed the title
bool to int casting issue in Python
4 participants
Description
When we need to check
if var == 1, wherevaris a bool type and 1 is an integer type, we need to do an unsigned extension of i1 since a signed extension of bool True will result in -1 rather than 1.Also, add the type casting logic to
NotEqas well since it should be equivalent toEq.Resolve: #1620