gh-127544: Ensure that AF_DECnet is part of socket.AddressFamily, when available #127548
Conversation
| @@ -0,0 +1,2 @@ | |||
| Ensure that ```AF_DECnet``` is part of the :class:`!socket.AddressFamily` | |||
There was a problem hiding this comment.
I'm not sure why :class:`socket.AddressFamily` didn't work.
| Ensure that ```AF_DECnet``` is part of the :class:`!socket.AddressFamily` | ||
| enum, when available. |
There was a problem hiding this comment.
| Ensure that ```AF_DECnet``` is part of the :class:`!socket.AddressFamily` | |
| enum, when available. | |
| Ensure that ``AF_DECnet`` is part of the :class:`!socket.AddressFamily` | |
| enum. |
There was a problem hiding this comment.
Why remove "when available"? It's in an ifdef in socketmodule.c, so CPython already hedges that it might not be present on any given system.
There was a problem hiding this comment.
I think it's a bit redundant, and when it's available, it'll appear in AddressFamily, otherwise it won't. If it isn't modified, it seems that it must be in AddressFamily, rather than changing with the os environment.
There was a problem hiding this comment.
That's the opposite of how I'd read it. "when available" at the end meaning "when it's available in the OS environment, ensure that it's part of AddressFamily". Without that, I'd think that "ensure that it's in AddressFamily" means that we always ensure it's in AddressFamily, whether or not it's available from the OS, or that we think it's universally available so not worth mentioning that part.
|
An alternate solution would be to keep |
This MR removes the requirement that AddressFamily members are all-uppercase, in order to include
AF_DECnet. I added a test to validate that it's part of the enum if the value is present in_socket.