Re: [PATCH] Implement SRFI-111 Boxes

[Ludovic Courtès]

Andy Wingo <address@hidden> skribis:

> On Tue 25 Mar 2014 18:01, address@hidden (Ludovic Courtès) writes:
>
>> Tu put it differently, I don’t think it would buy us anything to make
>> variable SRFI-111 boxes.
>
> Dunno; variables are slightly cheaper than records.  Their type checks
> are easier and they take less memory.  
>
>> However, it could perhaps break code
>
> In what way?

‘variable?’ would suddenly match any SRFI-111 box.

>> and would not allow for a separate type printer, which is useful.
>
> Not sure what this point is, as the external representation was not
> specified in SRFI-111.

I mean this:

  #<variable 15cb830 value: #<procedure + (#:optional _ _ . _)>>

vs.

  #<box 206f8c0 value: #<procedure + (#:optional _ _ . _)>>

I find it convenient that variables are distinguished.

> Perhaps we are miscommunicating -- I would change what we currently call
> "variables" to be "boxes".  WDYT?

You mean just the name, or making them SRFI-111 boxes?

Anyway, I don’t feel strongly about this particular point, but I think
we’ve been deprecating a lot lately.

Ludo’.