Re: compatibility package for ANSI CL?

Francois-Rene Rideau wrote:

> On Fri, Apr 05, 2002 at 10:13:00AM -0500, reti@xxxxxxxxxx wrote:
>
> >     While trying to run code, I've found that things like READ-SEQUENCE,
> >     DECLAIM, (LAMBDA ...) without #', and more, are lacking.
> Weird:
> 1) I implemented READ-SEQUENCE based on
>         (scl:send stream :string-in nil ar start end)
>  but found that it was reading files line-by-line, which is *extremely*
>  inefficient. :read-bytes isn't accepted by the NFS file I'm working on
>  either. I guess I ought to try to peruse the sources for Zmacs to see how
>  it ultimately does things, but I wonder what is the Right Thing(tm).

the "Right Thing" here is probably ansi-cl:read-block, which may not be
implemented yet.

I've never actually used the official ANSI-CL stuff which wasn't in Steele 2, 
far
as I can recall...


> 2) when my defpackage was :use'ing FUTURE-COMMON-LISP,
>  #'(lambda ...) expanded as (cl:function (future-common-lisp:lambda ...))
>  which didn't work. Using Syntax: Common-Lisp and package Common-Lisp
>  instead of Syntax: ANSI-Common-Lisp and Common-Lisp or Future-Common-Lisp
>  solved the problem.

I hate all these screwy package use, import and ordering problems...


> 3) Using Common-Lisp, just defmacro'ing declaim to future-common-lisp:declaim
>  did the trick.

the "Right Thing" here is more probably

(setf (function 'declaim) (function 'fcl:declaim))

    or

(defpackage YOUR-PKG (:import (fcl:declaim)) ...)

or something equivalent to that, I haven't done it in a long time, but I know I
had to do the same thing some time in the past.


> Probably unrelated to Genera, the Mac IIfx host seems to have major
> instability problems, although NiftyTelnet SSH runs happily.
> The floppy when in write mode may destroy floppies; more importantly
> I think the RAM might be bad: files imported via netatalk might
> be partly garbled with the recurring pattern DB 6D B6, IE will crash badly
> after a few minutes, the chooser itself might become instable, etc.
> Can anyone recommend any diagnosis tool or method?

replacement floppy drives can be had for about US$20 now. RAM for that machine 
is
probably harder to find, it had special SIMM sticks. I'd be willing to be that 
the
correct diagnostic software is going to be very hard to find...


> >     Trying to get the Lisp lore without a guru at hand is not easy.
> >     Did you bootstrap your learning by being in contact with gurus,
> >     or did you have tips to learn without a guru at hand?
> >
> > The main tip I have is "m-. is your friend", i.e. look at the source.
> > A corollary is that the c-m-<suspend>, the debugger, trap on call and
> > trap on exit, (si:enable-who-calls :all) and m-x List callers are your
> > friends too.

> Cool. It nice to have such friends. I'll have to tame their companions,
> too (e.g. I suppose there's a way to go to the next definition of save-buffer,
> after I m-. it and zmacs tells me there's 1 more definition of it).
> In any case, thanks a lot for your tips. From what I understand, interacting
> with the beast, rather than merely reading the docs, is how you can really
> learn it.

absolutely, there's A LOT more available than the docs have *EVER* actually
described.

I think it was meta-, which jumps to the next one, but only after you've done
meta-.

 -- clint, former TI Explorer guru