Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!pasteur!jwz@teak.berkeley.edu
From: jwz@teak.berkeley.edu (Jamie Zawinski)
Newsgroups: comp.sys.ti.explorer
Subject: RE: more microexp-mac connection
Message-ID: <18540@pasteur.Berkeley.EDU>
Date: 19 Oct 89 17:58:00 GMT
Sender: news@pasteur.Berkeley.EDU
Lines: 61
BCC:         Jamie Zawinski <jwz@teak.berkeley.edu>
In-reply-to: <8910180954.AA02523@lri.lri.fr>

> - Convert a pure MacIntosh font to an explorer font (ex: chicago-12)
>   I want to be able to save this font into an XLD file that can be
>   read on the explorer)

Clint mentioned his code which does this; you can get it by anonymous FTP
to spice.cs.cmu.edu (128.2.254.139) in the file 
/usr/jwz/public/mac-font-convert.lisp.  I had some trouble using this code,
so I found an easier way: a whole lot of macintosh fonts are distributed 
with X11r3 in BDF font format; I have code which reads this format in
the above-mentioned directory.  Also, in the fonts/ subdirectory are 
compressed tar files of Explorer XLD versions of over a hundred mac fonts,
and all of the Adobe 75DPI fonts which came with X11r3.

> - Convert a MacIntosh bitmap (i.e. MacPaint file) into an explorer
>   bitmap (i.e. loadable using w:read-bit-array-file)

I have code for reading Amiga IFF files.  DeluxePaint on the Macintosh
produces files in this format; if you have a converter to DPaint's 
format, my code will probably read it.  It will read monochrome or color
images, but can't do dithering (anyone want to implement that?)

> - A final question. We do not have release 6 yet, and I wonder how
>   different the interface building tools are with respect to release
>   5.

TI just sent us release 6 last week (I'd had it for a couple of months 
before that, but I ain't sayin' how :-))  I encountered one problem of the 
type you fear; I have code which defines a new flavor of Lisp listener
in which everything printed with PRIN1 or PRINC is mousable - left returns
the object, right brings up a menu of operations on it.  This worked great
in Rel4, I just specialized the :PRINT method.  But now, all of that streams
stuff has been re-implemented in CLOS, so I had to write some CLOS methods
to get it working again.

This was hard, because the CLOS implementation of PRINT-OBJECT specializes
on the first argument (doing different things for vectors, symbols, etc).
I wanted to specialize on the second argument, the stream.  Ok, no problem
there, right?  Well, yes, there is.  I can't just say

(defmethod ticlos:print-object :around ((object t) (stream my-window))
  ... )

because if something is printed to *STANDARD-OUTPUT* (a common occurrence)
*STANDARD-OUTPUT* will hold a SYNONYM-STREAM instead of a MY-WINDOW.

So then I thought, ok, wrap print-object with some code that will replace
the syn-stream with what it points to, and then reinvoke.  That will let
the method-dispatching on the second argument work.

How do you tell if something is a synonym stream?  The only way I could
decide was (AND (SYMBOLP STREAM) (FBOUNDP STREAM)) which is pretty evil.

> Is there a real support for multi-windows overlaping each others or are 
> we still in the old philosophy of one window at a time?

Get the KSL code from Rice and Acuff.  It's really great stuff.  Right now
I'm typing this in a Zmacs taking up half of the screen, while on the other
half, a desktop is visible, with icons representing other application
windows.

			-- Jamie