[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: vm merged to master
From: |
Andy Wingo |
Subject: |
Re: vm merged to master |
Date: |
Mon, 16 Mar 2009 23:54:48 +0100 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/23.0.91 (gnu/linux) |
Heya Ludo,
On Sat 14 Mar 2009 15:36, address@hidden (Ludovic Courtès) writes:
>> `(procedure-property apply 'arity)' needs some work
>
> Right, that won't work if `apply' is a program. The question is: what
> should we do with procedure properties?
You jump questions ;)
1) Procedure properties do work on programs. They just aren't
serialized to disk. But some serializations set properties on the
resulting procedure -- e.g. define*.
2) Programs do have properties -- program-properties, i think the
accessor is.
3) Should arity be accessed via procedure-properties? I think not.
Logically properties are cached values, not computed values, whereas
arity can be computed.
3.5) Should arity even be represented like that? The recent addition to
(ice-9 session), procedure-arguments, is another take on that
question.
4) For some reason I like numbered lists recently? Please take it as
some way of organizing my thoughts, it's not intended to be pedantic
;)
> Before the VM, [procedure-properties] could be used both with subrs
> and Scheme closures.
They still can.
> - the `object-ref' limitation to 256 objects;
We spoke of this on IRC, for those that weren't there. Actually, for
those that don't know: #guile is a friendly place. We actually have a
night shift going during the week, if you drop in between 21h and 24h
CET, people are likely hacking away :-)
> - `current-reader' not taken into account;
I want to fix this soon.
> - circular references among modules (previously worked around via
> autoloads) not supported---but this is a good thing, it's an
> opportunity to remove kludges. :-)
Is this really the case? Can you make a test case?
> I'll eventually proceed with the merge in the BDW-GC branches.
Great, I'm looking forward to that :-))
Cheers,
Andy
--
http://wingolog.org/