Re: [PATCH] * etc/NEWS: Document incompatibilities introduced by record types.

[Stefan Monnier]

> My suggestion would be to either at least document this breaking change
> (this patch), or forbid such records,

Hello?  We've been through this discussion enough times, I believe.
I'm absolutely firmly completely totally opposed to such a check in
`make-record`.

Yes, a record of type `integer` or `hash-table` is not allowed.
No, we don't want to enforce this with a check in `make-record`.

> or change the return value in a non-breaking way (such as using
> '(record integer) and "#r(integer)").

Other parts of Emacs rely on the fact that (eq (type-of X) (type-of X)),
so if you want to return cons cells, you'll need to hash-cons them or
something.  And the use of `integer` as record type is incorrect in any
case, so it would add complexity only to handle a situation that's
already incorrect.

You're making a mountain of a mole-hill, really.


        Stefan