Backport #7936: backport r39449

• Status: Rejected
• Assignee: nobu (Nobuyoshi Nakada)

[zzak (zzak _)]

[mame (Yusuke Endoh)]

• Status changed from Open to Assigned
• Assignee changed from mame (Yusuke Endoh) to nagachika (Tomoyuki Chikanaga)

This is not a typo fix or format improvement, but includes (very small) semantic change.
Let's consider after the patch level release after p0.

--
Yusuke Endoh [e-mail]

[nagachika (Tomoyuki Chikanaga)]

I think passing String to these methods is meaningful because it is a way to avoid unnecessary Symbol allocation. So I'll merge r39449.

mame san, do you have a concern?

[nobu (Nobuyoshi Nakada)]

Yes, it's meaningful, but the documents are inaccurate on that point.
Any *_defined? methods don't convert the argument to a Symbol always.

I don't think the distinguishment between symbol and string here is important enough.
And the name symbol/string seems less meaningful.
I'd suggest to change the argument names and mention that it can be a symbol or a string.

[zzak (zzak _)]

nobu (Nobuyoshi Nakada) wrote:

Yes, it's meaningful, but the documents are inaccurate on that point.

Thank you for your feedback, always appreciated nobu-san.

I'd suggest to change the argument names and mention that it can be a symbol or a string.

Do you mean for *_defined? methods or for all methods mentioned in this patch?

[nagachika (Tomoyuki Chikanaga)]

• Status changed from Assigned to Feedback
• Assignee changed from nagachika (Tomoyuki Chikanaga) to nobu (Nobuyoshi Nakada)

nobu, could you clarify about zzak's question?

[nobu (Nobuyoshi Nakada)]

I meant for all methods.

Sorry for late reply.

[zzak (zzak _)]

@nobu (Nobuyoshi Nakada) np, you mean something like:

attr_accessor(string or symbol)

or

attr(string or symbol)

[naruse (Yui NARUSE)]

• Status changed from Feedback to Rejected