Re: [Chicken-hackers] [PATCH] Fix #1581 by allowing unqualified record names

[felix . winkelmann]

> address@hidden wrote:
> > Can anybody give details about the coops problem?
>
> From what I can remember, the main problem is not in coops itself, but
> in record-variants, that uses low-level ##sys# procedures to define
> records and doesn’t use the right tags (which don’t work with
> define-record-printer and in turn prevent print-object methods from
> working as intended).

Ok, thanks for the information.

>
> I think with the module prefix, any egg using the low-level structure
> procedures have to be updated.
>
> If I remember correctly, at SaarCHICKEN we pretty much settled on the
> solution of using the record name variable containing the correct tag,
> which is the most correct and elegant solution, and should fix every
> one of these problems, not just `record-instance?`

I don't disagree, we just need certain eggs to be in a working state,
and I'm trying to assess the situation. If one is using the low-level accessors
than he or she should simply know what they're doing. But I consider coops
(as limited as it is) as sort of fundamental, so it would be good if we find
a solution to make the basic stuff usable.

>From a quick scan I see no reason why record-variants shouldn't be able
to use the tag variable instead of the quoted variant name. I'll investigate...


felix