Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!mcnc!rti!xyzzy!kan
From: kan@dg-rtp.dg.com (Victor Kan)
Newsgroups: comp.lang.lisp.x
Subject: Re: exploratory system for unix system calls/facilities
Summary: partially what I'm looking for, Byte vol 13, #11 IBM special edition
Message-ID: <1852@xyzzy.UUCP>
Date: 15 Oct 89 00:27:48 GMT
References: <1753@xyzzy.UUCP> <4139@hplabsz.HPL.HP.COM>
Sender: usenet@xyzzy.UUCP
Reply-To: kan@mutley.dg.com ()
Organization: Data General Corporation, Research Triangle Park, NC
Lines: 86

In article <4139@hplabsz.HPL.HP.COM> mayer@hplabs.hp.com (Niels Mayer) writes:
>In article <1753@xyzzy.UUCP> kan@tom.dg.com () writes:
>>Does anyone know of an interactive environment for playing around with
>>all those wonderful Unix system calls?
>>
>I've added a few "easy" unix system (e.g. system(3), popen(3) etc) calls to
>xlisp in my WINTERP X11/Motif application prototyping, development and
>delivery environment. Furthermore, in WINTERP, the xlisp read-eval-loop has
>been transformed into a lisp server allowing for clients to send
>lisp-expressions from other applications distributed across the network. Is
>this what you mean by a simple client/server? Haven't touched (nor do i
>want to get my hands dirty with) shared memory, semaphores, etc. But then
>again, I'm building this system as a platform for our research and
>experiments in multimedia computing and groupware, so our aims might be
>different.

When I say client/server, I'm not refering to a distributed XLISP system,
but client/server prototypes implemented in XLISP.  Hopefully, I could
debug a protocol in the XLISP interpreter rather than a C compiler.

>
>Mostly, I've been concentrating on making a clean "object-oriented"
>interface to the Motif widgets. By the time I release WINTERP, it may also
>have an interface to the Xtoolkit's XtAddInput() routine which will provide
>for a simple way of having WINTERP maintain open connections to multiple
>network clients; intefacing the XtAddWorkProc() routine should allow for
>simple internal process scheduling.
>

Has it been tricky using XLISP rather than C/C++ to interface with the
X library (excuse me, Motif Widgets :-), or has the joy of lisp offset
the headaches?  

>I'd certainly be interested in getting any enhancements to xlisp for
>supporting some of the uglier aspects of unix hacking in a cleaner fashion.
>My winterp code should be publically available by November 24th, and will
>hopefully be included on the X11r4 contrib tape.
>
>>I recall that Byte had a description of XLisp extensions on OS/2 that 
>>allowed you to use system services interactively.
>
>I haven't seen this. Which issue?

I've been looking for the issue for weeks, but finally found it today.
It's in the Byte IBM Special Edition for Fall 1988 (volume 13, #11).
To summarize, C library functions are called with the c-call function.
C style structures are handled like this:

/* from their example */
struct FileFindBuf {
	unsigned create_date;
	unsigned create_time;
	unsigned access_date;
.
.
.
}

becomes:

(define FileFindBuf
	'((word create_date)
	  (word create_time)
	  (word access_date)
.
.
.
))

Unfortunately, the source for the extentions of XLISP to OS2XLISP were not
published in the issue.  I assume it's available from the usual Byte
source sources.

>
>-------------------------------------------------------------------------------
>	    Niels Mayer -- hplabs!mayer -- mayer@hplabs.hp.com
>		  Human-Computer Interaction Department
>		       Hewlett-Packard Laboratories
>			      Palo Alto, CA.
>				   *
>


| Victor Kan               | I speak only for myself.               |  ***
| Data General Corporation | Edito cum Emacs, ergo sum.             | ****
| 62 T.W. Alexander Drive  | Columbia Lions Win, 9 October 1988 for | **** %%%%
| RTP, NC  27709           | a record of 1-44.  Way to go, Lions!   |  *** %%%