Improve handling of bitfields in ctypes #1441
Conversation
BCSharp
changed the title
BCSharp
force-pushed
the
BitField_ctypes
branch
2 times, most recently
from
22453b3
to
5d07702
Compare
| } | ||
| } | ||
| } else { | ||
| throw PythonOps.TypeErrorForTypeMismatch("int", value); |
There was a problem hiding this comment.
I think this bit is unreachable so you can throw InvalidOperationException?
There was a problem hiding this comment.
Right. It's a left-over from a pre-3.8 code version.
| value = (BigInteger)bits; | ||
| } | ||
|
|
||
| } else if (value is bool bVal) { |
There was a problem hiding this comment.
Is this reachable or bool will be caught by the IsBoolType check above?
There was a problem hiding this comment.
No, not reachable. Again a leftover from a previous version.
| @@ -50,6 +50,10 @@ public abstract class _Structure : CData { | |||
| INativeType nativeType = NativeType; | |||
| StructType st = (StructType)nativeType; | |||
|
|
|||
| if (args is null) { // None passed as a sole argument | |||
There was a problem hiding this comment.
Interesting, None gives null here? Would adding the NotNoneAttribute solve this?
There was a problem hiding this comment.
Yes, None gives null. This is the consequence that the DLR allows to bind to the param object[] array directly (just like C#). We have discussed it before (IronLanguages/dlr#249). Although then I was arguing against it, I now think that for params arrays it will be better to keep the behaviour as it is now.
But indeed, [NotNone] prevents it from happening, pushing null into the params array. Good insight! It is an elegant solution.
|
Thanks for the review! |
Addressed #1439 and a few other issues that caught my attention. The most interesting is a fix for python/cpython#90914, which was reported just 3 months ago and is still not resolved. Funny thought that IronPython on one point will be ahead of CPython. Not that this functionality is used anywhere in the wild, since it is not working yet in CPython.