[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Gcl-devel] Re: Can't subclass STANDARD-CLASS

From: Dan Debertin
Subject: [Gcl-devel] Re: Can't subclass STANDARD-CLASS
Date: Tue, 14 Feb 2006 05:28:36 +0000 (UTC)
User-agent: Loom/3.14 (http://gmane.org/)

[[ Following-up my own message with more info ... ]]

Dan Debertin <airboss <at> nodewarrior.org> writes:

> (defclass test-metaclass (standard-class)
> ())
> Error in MAKE-INSTANCE [or a callee]: MAKE-INSTANCE [or a callee] requires 
> less
> than three arguments.

[ ... etc ... ]

Some progress -- after chasing callees of MAKE-INSTANCE down several blind
alleys looking for functions getting called with the wrong number of args, two
things became apparent:

1. It would be a LOT easier to debug this if the backtrace could extend back
into what is already compiled into the image. That appears to be impossible, but
I'd like to be wrong here. 

Failing that, is there a way to re-load all of PCL from source so a backtrace
will be more useful? In what order should the gcl_pcl_*.lisp files be loaded in
order to accomplish this?

2. On a lark, I loaded the saved_pcl_gcl image (I was using saved_ansi_gcl
before) and tried the same thing. That image will define subclasses under
STANDARD-CLASS with no issues. Why/how is PCL different from the ANSI image?



reply via email to

[Prev in Thread] Current Thread [Next in Thread]