We can tolerate PyPy now... #1139
Conversation
| @@ -29,7 +29,7 @@ def eval_N( | |||
| expression: BaseElement, | |||
| evaluation: Evaluation, | |||
| prec: BaseElement = SymbolMachinePrecision, | |||
| ) -> BaseElement: | |||
There was a problem hiding this comment.
This type mismatch and the one above were noticed when looking at another bug we have.
There is pollution to values of some non-Python data. (Probably some structure is modifying a list in place).
A fix for the underlying problem will occur down the line.
| @@ -134,7 +134,7 @@ def quiet_f(*args): | |||
|
|
|||
|
|
|||
| def eval_ListPlot( | |||
| plot_groups: list, | |||
There was a problem hiding this comment.
Should it not be better to ensure using one type or the other?
There was a problem hiding this comment.
Hmm...
plot_groups is sometimes the result of expr.to_python(). For a ListExpression this is a list.
In the case of constant data we have a tuple; here, this is synonymous with an immutable list.
I think plot_groups should be a tuple - immutable. However, I don't think we can change the return type of to_python() and I don't think we can change it to tuple on return without first detecting that it is a list. Otherwise, we would be wrapping a scalar into a tuple.
So, while I agree it would be better to ensure a tuple is passed in, at this stage, I am not sure I want take the risk of converting a list to tuple right now.
I have made note of this issue though in commit 016f6be
There was a problem hiding this comment.
OK, I think it would be enough by now.
Fixes #661