dataclasses.Field and dataclasses._DataclassParams are missing spaces in their reprs between their attributes

[osvathbotond]

Bug report

Bug description:

For example, the following code

import dataclasses


@dataclasses.dataclass
class User:
    name: str
    age: int

print(repr(dataclasses.fields(User)))

will output

(Field(name='name',type=<class 'str'>,default=<dataclasses._MISSING_TYPE object at 0x70d5c2a11550>,default_factory=<dataclasses._MISSING_TYPE object at 0x70d5c2a11550>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),kw_only=False,_field_type=_FIELD), Field(name='age',type=<class 'int'>,default=<dataclasses._MISSING_TYPE object at 0x70d5c2a11550>,default_factory=<dataclasses._MISSING_TYPE object at 0x70d5c2a11550>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),kw_only=False,_field_type=_FIELD))

CPython versions tested on:

3.13

Operating systems tested on:

Linux

Linked PRs

• gh-137714: Improving the repr of dataclasses.Field and dataclasses._DataclassParams #137715

[ericvsmith]

Is this causing some actual problem? I'm hesitant to change it if it's just a cosmetic thing, because no doubt people are depending on the existing repr.

[osvathbotond]

@ericvsmith Thank you for taking a look at it! I don't think it causes any problem, I just came across it by fiddling around with dataclasses. Could you give me a situation where people would depend on it? Do you expect people parsing the repr for attribute values instead of getting them directly from the object itself?

[ericvsmith]

Doctests, for one. And for all I know people have formatted fix-width output that needs the repr to be this length. One thing I've learned in the last 20 years is that people depend on everything! And unless there's a good reason, I'd rather not change it.

[osvathbotond]

@ericvsmith Thank you for your time and feedback, I really appreciate it! I agree with you.