bug#31715: cl-incf and cl-decf error out when passed a nil-valued variable as 'offset'

["Clément Pit-Claudel" <clement.pitclaudel@live.com>]

[-- Attachment #1.1: Type: text/plain, Size: 1114 bytes --]

On 2018-06-05 19:26, Noam Postavsky wrote:
> Clément Pit-Claudel <clement.pitclaudel@live.com> writes:
> 
>> I'm not sure what to make of the result with SBCL and CLisp, since
>> (incf x nil) also fails in both of them (whereas it works for us,
>> since we can't distinguish nil and unspecified).
> 
> Oh, huh, I didn't think to check that case.  Maybe we should just change
> cl-incf to disintguish between nil and unspecified then.

Hmm, neat trick! Wouldn't that be backwards-incompatible, though?
Plus, it's not common for ELisp functions to distinguish between unspecified and nil, and changing cl-incf that way would make it less convenient to wrap in other macros.  If it doesn't had runtime costs, I'd be in favor of going the other route, and making sure that (cl-incf x y) adds 1 even when y is bound to nil.

>>> I don't think those are great examples of macros to emulate.
>>
>> Agreed, I was just collecting other examples, both in support and against my point.
> 
> Yeah, I just meant we can't really use those examples either to support
> or argue against your point.

OK :)


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]