NameError for HyList, HyDict ...

[nalssee]

Evaluating HyDict raises a NameError but once I evaluate '{}, it is defined in the env.

So I had to write like (instance? (type '{}) expr) instead of (instance? HyDict expr).

All the other HyObjects are the same.

I'm not sure if it is an intended behaviour or a bug.

[refi64]

Each of the classes in within a module in hy.models:

• HyDict is in hy.models.dict.
• HyList is in hy.models.list.
• etc.

[gilch]

Hy doesn't always import things until it needs them. You can use hy --spy to see what's going on. (--spy attempts to show you the Python code that would generate the same AST Hy has generated.)

=> HyDict
HyDict
Traceback (most recent call last):
  File "<input>", line 1, in <module>
NameError: name 'HyDict' is not defined
=> '{}
from hy import HyDict
HyDict([])
{}
=> HyDict
HyDict
<class 'hy.models.dict.HyDict'>

The key line is from hy import HyDict. You can do this yourself if you prefer.

[refi64]

Odd behavior, but intentional. I'll just close this for now.

[Kodiologist]

Actually, I'm inclined to think this a bug; if we need to import a name that isn't a Hy builtin, it should be imported with a gensym. But it would make sense to make HyList etc. builtins.

[tuturto]

Indeed, would make sense to have HyDict and friends available without explicitly importing them.

[gilch]

Let's add the Hy models to core then.

[Kodiologist]

Gilch (#1360 (comment)) mentioned that HyCons isn't auto-imported in the way that the other Hy model types currently are. I'm guessing that's just an oversight.