jscheme-devel — Developer's list

Re: [Jscheme-devel] Re: High performance JScheme servlets?

[Geoffrey K. <ge...@kn...>]

On Sep 28, 2004, at 4:34 PM, Ken Anderson wrote:
> Could we get a servlet that only reads the file once?  Or maybe even 
> never - it calls a function that its already loaded?

I think it should be an option.  load-always or load-once, so to speak.

At the customer site where I've been using JScheme servlets to demo 
screen designs and change them on the spot, it's been very nice that 
little tweaks to my html.scm and defs.scm "libraries" get reflected 
immediately in the browser, and fellow developers were impressed that I 
never had to stop/restart the Jetty server.

I can see the value of load-once too, so long as there is a way to dump 
("forget") something in memory dynamically so a new version can be 
loaded.

Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk

[Jscheme-devel] Re: High performance JScheme servlets?

[Timothy J. H. <ti...@cs...>]

On Sep 28, 2004, at 4:34 PM, Ken Anderson wrote:

> Tim,
>
> Your various servlets always read and evaluate a file, which is fine 
> for your course and development, but what about performance?
>
> Could we get a servlet that only reads the file once?  Or maybe even 
> never - it calls a function that its already loaded?

I haven't though much about it, but its a good idea.

A partial solution would be to write servlets as follows:
(servlet (a b c ... z)
   (use-module "/full/path/to/file/F.scm")
   (F a b c ... z)
)

The server would still have to read and parse the file, but it would be
a relatively small file.

One could also skip even this small amount of parsing by using a cache
to see if the file has been read before... (This would go in 
WEB-INF/scheme/scheme.sss)

I haven't tried either of these, but it would be interesting to see how 
much they improve
performance. (I've found httperf to be a nice tool for measuring 
servlet performance....)

---Tim---


> k
>

Re: [Jscheme-devel] passing parameters to JScheme servlets

[Geoffrey K. <ge...@kn...>]

Ken and Tim  -- Thanks for your suggestions.  What I discovered was 
close to what Ken suggested (hidden parameters), but I can see utility 
for Tim suggested (session parameters).  Here is the test code I used.

Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk

;;;;; send.servlet

(servlet
  ()
  (define (buttonSend buttonValue action att val)
    (sa-form (catt (s-method "post") (s-action action))
             (s-input (s-type "submit") (s-value buttonValue))
             (s-input (s-name att) (s-type "hidden") (s-value val))))
  (load "webapp/jscheme/test/html.scm")
  (buttonSend "Send my name" "recv.servlet" "firstName" "Geoffrey"))

;;;;; recv.servlet

(servlet
  ()
  (load "webapp/jscheme/test/html.scm")
  (let ((name (.getParameter request "firstName")))
    (s-html
     (s-p "Your first name is: " name)
     (go-button "again" "send.servlet"))))

[Jscheme-devel] High performance JScheme servlets?

[Ken A. <kan...@bb...>]

Tim,

Your various servlets always read and evaluate a file, which is fine for your course and development, but what about performance?

Could we get a servlet that only reads the file once?  Or maybe even never - it calls a function that its already loaded?
k

Re: [Jscheme-devel] passing parameters to JScheme servlets

[Timothy J. H. <ti...@cs...>]

On Sep 28, 2004, at 2:12 PM, Geoffrey Knauth wrote:

> How do I pass parameters between JScheme servlets?
There are several techniques ...
As Ken suggests, hidden parameters is a standard one
>    <input name="abc" type="hidden" value="xyz"> ..


Another approach is to use session parameters which
are stored on the server and accessible to servlets in the same session.
I do this using the following two procedures for storing and accessing
session values (which is in the jscheme webapp from CVS in
WEB-INF/scheme/servletlib/session.scm

> ;; storing session data
>  (define (session-set request N V)
>    (.setAttribute (.getSession request) N V))
>
>  (define (session-get request N Default)
>   (let ((x (.getAttribute (.getSession request) N)))
>    (if (equal? x #null) Default x)))
This approach is good for passwords where you get them once
from the user and the store them in a session attribute.

Does this help??

You could probably also "call a servlet" and get its return value by 
doing a read on a URL...
>
>   (define result
>        (http-read
>            
> "http://my.company.com:8080/jscheme/my.servlet?a=1&b=2&c=3&d=hi"))
>

but this can expose you to the danger of infinite servlet loops where 
servlet A calls servlet B
who calls servlet C, etc.

The http-read procedure can be defined as follows (but there might be a 
more elegant solution....)

> (define (http-read url-string)
>    ;; this reads a string from a given url
>    (define url (java.net.URL. url-string))
>    (define http-con (.openConnection url))
>    (define buf (java.lang.reflect.Array.newInstance char.class 65536))
>
>    (define (drain-from-in in) ;; read input lines into a list in 
> reverse order
>     (let loop ((n (.read in buf 0 65535))
>                (L ()))
>          (if (equal? n -1)
>              L
>              (let ((z (java.lang.String. buf 0 n)))
>                (loop (.read in buf 0 65535) (cons z L))))))
>
>    (define (strip-out-header L) ; L is a reversed list of lines from 
> an http file
>        (let loop ((L L) (R ()))
>          (cond ((null? L) R)
>                ((equal? (first L) "") R)
>                (else (loop (rest L) (cons (first L) R))))))
>
>    (define (list->string L)
>      (let loop ((L L) (z (java.lang.StringBuffer.)))
>           (if (null? L)
>               (.toString z)
>               (loop (rest L) (.append (.append z (first L)) "\n")))))
>
>    (let (
>           (in (java.io.BufferedReader.
>                (java.io.InputStreamReader.
>                 (.getInputStream http-con)))))
>      (let* ((result (drain-from-in in)))
>         (.disconnect http-con)
>         (list->string (strip-out-header result))
>     )))


---Tim---



>
> Geoffrey
> --
> Geoffrey S. Knauth | http://knauth.org/gsk
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: IT Product Guide on 
> ITManagersJournal
> Use IT products in your business? Tell us what you think of them. Give 
> us
> Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out 
> more
> http://productguide.itmanagersjournal.com/guidepromo.tmpl
> _______________________________________________
> Jscheme-devel mailing list
> Jsc...@li...
> https://lists.sourceforge.net/lists/listinfo/jscheme-devel

Re: [Jscheme-devel] passing parameters to JScheme servlets

[Ken A. <kan...@bb...>]

Tim can probably answer this better than i can.  Also a bit more context of what you want to do would help.

Can you put the information in a hidden field in a form?

At 02:12 PM 9/28/2004 -0400, Geoffrey Knauth wrote:
>How do I pass parameters between JScheme servlets?
>
>Geoffrey
>--
>Geoffrey S. Knauth | http://knauth.org/gsk
>
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: IT Product Guide on ITManagersJournal
>Use IT products in your business? Tell us what you think of them. Give us
>Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out more
>http://productguide.itmanagersjournal.com/guidepromo.tmpl
>_______________________________________________
>Jscheme-devel mailing list
>Jsc...@li...
>https://lists.sourceforge.net/lists/listinfo/jscheme-devel

[Jscheme-devel] passing parameters to JScheme servlets

[Geoffrey K. <ge...@kn...>]

How do I pass parameters between JScheme servlets?

Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk

[Jscheme-devel] / => \ in [Win] bin\make.bat

[Geoffrey K. <ge...@kn...>]

To get bin/make.bat to work on Windows (our tech lead's machine), I had 
to change all the / chars to \ chars.  Subsequently, he was very 
impressed with JScheme's capabilities.

Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk

[Jscheme-devel] FAQ entry: (* 256 256 256 256) --> 0

[Timothy J. H. <ti...@cs...>]

On Aug 5, 2004, at 12:57 PM, Geoffrey Knauth wrote:

> Maybe this is a FAQ.  I guess I was surprised I got 0 instead of #f 
> and/or an exception of some kind.
>
> $ java -cp lib/jscheme.jar jscheme.REPL
> JScheme 7.1 (8/5/04 3:06 AM) http://jscheme.sourceforge.net
> > (* 256 256 256 256)
> 0
> > (* 256 256 256)
> 16777216
> > (* 256 256 256 128)
> -2147483648
> >

I guess we need a FAQ for JScheme and questions about numerics need to 
be there...

---------------------

Q: Why does integer arithemetic give numerically incorrect results in 
JScheme?

A: JScheme numbers and operators are taken from Java so, for example,
     256 in an "int" and hence raising 256 to the fourth power gives 
zero in JScheme

  >      (display (* 256 256 256 256))
   0

  This gives zero because it gives the same result as a corresponding 
Java program,
  such as the following:

  public class Demo {
     public static void main(String[] args) {
        int x = 256;
         System.out.println(x * x * x * x);
     }
    }

Java integer arithmetic is does not throw exceptions on overflow and
just returns the lower order 32 bits of the answer, hence you get 0 for

If you want the Scheme numeric tower (which is an optional part of 
R4RS) you
can import the SISC tower and use the siscnum.scm module described here:
      http://localhost:8080/jscheme/src/using/package.html
However, if you use this you must explicity convert SISC numbers to 
Java numbers
before using those numbers in javadot calls.

---Tim---



>
> Geoffrey
> --
> Geoffrey S. Knauth | http://knauth.org/gsk
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by OSTG. Have you noticed the changes on
> Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now,
> one more big change to announce. We are now OSTG- Open Source 
> Technology
> Group. Come see the changes on the new OSTG site. www.ostg.com
> _______________________________________________
> Jscheme-devel mailing list
> Jsc...@li...
> https://lists.sourceforge.net/lists/listinfo/jscheme-devel

[Jscheme-devel] (* 256 256 256 256)

[Geoffrey K. <ge...@kn...>]

Maybe this is a FAQ.  I guess I was surprised I got 0 instead of #f 
and/or an exception of some kind.

$ java -cp lib/jscheme.jar jscheme.REPL
JScheme 7.1 (8/5/04 3:06 AM) http://jscheme.sourceforge.net
 > (* 256 256 256 256)
0
 > (* 256 256 256)
16777216
 > (* 256 256 256 128)
-2147483648
 >

Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk

Re: [Jscheme-devel] (import "yada.yada.Yada")

[Geoffrey K. <ge...@kn...>]

I did that too, with (addClasspathUrl JAR).  I was just looking for a 
way to start off with the right classpath.  Thanks.

Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk

On Jun 24, 2004, at 16:18, Ken Anderson wrote:

> java -jar lib/jscheme.jar should be fine, just grow the class path at 
> runtime, see elf/infest.scm for an example.

Re: [Jscheme-devel] (import "yada.yada.Yada")

[Ken A. <kan...@bb...>]

java -jar lib/jscheme.jar should be fine, just grow the class path at runtime, see elf/infest.scm for an example.

At 03:46 PM 6/24/2004 -0400, Geoffrey Knauth wrote:
>OK, I see your logic.
>
>I wrote my note after I imported some classes from JAR files, and then to my surprise found, from (classpathURLs), that the classes were never actually loaded and various symbols remained undefined.  It turned out that happened because I launched JScheme via "-jar jscheme.jar" instead of "jscheme.REPL".
>
>I guess I'd call the `import' facility that requires classes to exist `import-exist' or `import-now'.
>
>Geoffrey
>--
>Geoffrey S. Knauth | http://knauth.org/gsk
>
>On Jun 24, 2004, at 15:35, Ken Anderson wrote:
>
>>I don't think it should check, though you can write a version that does.
>>Since i use JScheme to build my applications its normal for classes not to exist until after i've compiled them.

Re: [Jscheme-devel] (import "yada.yada.Yada")

[Geoffrey K. <ge...@kn...>]

OK, I see your logic.

I wrote my note after I imported some classes from JAR files, and then 
to my surprise found, from (classpathURLs), that the classes were never 
actually loaded and various symbols remained undefined.  It turned out 
that happened because I launched JScheme via "-jar jscheme.jar" instead 
of "jscheme.REPL".

I guess I'd call the `import' facility that requires classes to exist 
`import-exist' or `import-now'.

Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk

On Jun 24, 2004, at 15:35, Ken Anderson wrote:

> I don't think it should check, though you can write a version that 
> does.
> Since i use JScheme to build my applications its normal for classes 
> not to exist until after i've compiled them.

Re: [Jscheme-devel] (import "yada.yada.Yada")

[Ken A. <kan...@bb...>]

I don't think it should check, though you can write a version that does.
Since i use JScheme to build my applications its normal for classes not to exist until after i've compiled them.

k

At 10:24 PM 6/23/2004 -0400, Geoffrey Knauth wrote:
>I think I should get an error when an import is invalid or fails:
>
>> (import "java.io.File")
>#t   ; OK
>> (import "yada.yada.Yada")
>#t   ; Huh?
>
>Geoffrey
>--
>Geoffrey S. Knauth | http://knauth.org/gsk
>
>
>
>-------------------------------------------------------
>This SF.Net email sponsored by Black Hat Briefings & Training.
>Attend Black Hat Briefings & Training, Las Vegas July 24-29 - digital self defense, top technical experts, no vendor pitches, unmatched networking opportunities. Visit www.blackhat.com
>_______________________________________________
>Jscheme-devel mailing list
>Jsc...@li...
>https://lists.sourceforge.net/lists/listinfo/jscheme-devel

Re: [Jscheme-devel] src/jsint/Evaluator.java:85 cannot find Version class

[Ken A. <kan...@bb...>]

Somehow, i convinced myself that i could build JScheme without the Version.java.  I did not want it in CVS because it caused conflicts that caused the nightly build to fail.

I'll fix src/build/bootstrap.bat and change things so the version string is read as a resource.

k
At 07:54 AM 6/24/2004 -0400, Timothy John Hickey wrote:

>On Jun 24, 2004, at 4:08 AM, Geoffrey Knauth wrote:
>
>>On Jun 24, 2004, at 03:44, I wrote:
>>>For now I'll just make my own Version.java using what I see in (make-version).
>>
>>I take that back.  It seems Version.java was really supposed to go away, and (make-version) is out of date.
>The problem was that the new version (represented as a final static field in Version.java) was
>not making it into jsint/Evaluator.java
>
>
>I've fixed the problem for non-Windows machines by adding Version.java back into the CVS
>(so that the initial compilation of jscheme.REPL can generate an initial JScheme system),
>But I've modified the bootstrap code so that it first removes all class files before compiling that
>initial JScheme system. I think the problem was that the static final VERSION field was being
>compiled "in place" into the jsint/Evaluator.class file and hence could only be changed by
>removing the Evaluator.class file. This is a general problem that can arise so I think it is safer
>to remove all class files before beginning a build.
>
>I also modified the (make-version) code in build/jscheme-bootstrapl.scm so that it include the
>short version of the time as well as the date. This allows one to better distinguish between builds.
>I guess our policy is to have the date/time represent the build counter for any given version.
>Thus, rather than produce versions 7.1.1 7.1.2, 7.1.3, ... we produce versions 7.1 (date/time1)
>7.1(date/time2),...
>
>I'll work on the windows version src/build/bootstrap.bat when I get access to a windows machine
>so I can test it....
>
>
>>  Instead, I commented out line 85 of src/jsint/Evaluator.java:
>>
>>   //error.println("\n\n"+Version.VERSION+"\n\n");
>>
>>Now, src/build/bootstrap works.
>>
>>Personally, I'm happy enough with just a "> " prompt when I start up JScheme.  But maybe there should be a (help) or (jscheme-version) users could type if they're going through Emacs buffers and can't remember which ones are running JScheme / mzscheme / MIT/GNU Scheme / Kawa / Guile / SISC / ...
>
>You can always type jsint.Version.VERSION$
>but we probably do want a simpler way to access the version, e.g. (jscheme-version)
>(or how about    (JScheme.version) , i.e. a static method of jscheme.JScheme.java
>
>a (help) or (help x) might be nice too....
>(apropos x) goes a long way toward helping.... as does (describe x)
>
>---Tim---
>>
>>Geoffrey
>>--
>>Geoffrey S. Knauth | http://knauth.org/gsk
>>
>>
>>
>>-------------------------------------------------------
>>This SF.Net email sponsored by Black Hat Briefings & Training.
>>Attend Black Hat Briefings & Training, Las Vegas July 24-29 - digital self defense, top technical experts, no vendor pitches, unmatched networking opportunities. Visit www.blackhat.com
>>_______________________________________________
>>Jscheme-devel mailing list
>>Jsc...@li...
>>https://lists.sourceforge.net/lists/listinfo/jscheme-devel
>
>
>
>-------------------------------------------------------
>This SF.Net email sponsored by Black Hat Briefings & Training.
>Attend Black Hat Briefings & Training, Las Vegas July 24-29 - digital self defense, top technical experts, no vendor pitches, unmatched networking opportunities. Visit www.blackhat.com
>_______________________________________________
>Jscheme-devel mailing list
>Jsc...@li...
>https://lists.sourceforge.net/lists/listinfo/jscheme-devel

Re: [Jscheme-devel] src/jsint/Evaluator.java:85 cannot find Version class

[Timothy J. H. <tim...@ma...>]

On Jun 24, 2004, at 4:08 AM, Geoffrey Knauth wrote:

> On Jun 24, 2004, at 03:44, I wrote:
>> For now I'll just make my own Version.java using what I see in 
>> (make-version).
>
> I take that back.  It seems Version.java was really supposed to go 
> away, and (make-version) is out of date.
The problem was that the new version (represented as a final static 
field in Version.java) was
not making it into jsint/Evaluator.java


I've fixed the problem for non-Windows machines by adding Version.java 
back into the CVS
(so that the initial compilation of jscheme.REPL can generate an 
initial JScheme system),
But I've modified the bootstrap code so that it first removes all class 
files before compiling that
initial JScheme system. I think the problem was that the static final 
VERSION field was being
compiled "in place" into the jsint/Evaluator.class file and hence could 
only be changed by
removing the Evaluator.class file. This is a general problem that can 
arise so I think it is safer
to remove all class files before beginning a build.

I also modified the (make-version) code in build/jscheme-bootstrapl.scm 
so that it include the
short version of the time as well as the date. This allows one to 
better distinguish between builds.
I guess our policy is to have the date/time represent the build counter 
for any given version.
Thus, rather than produce versions 7.1.1 7.1.2, 7.1.3, ... we produce 
versions 7.1 (date/time1)
7.1(date/time2),...

I'll work on the windows version src/build/bootstrap.bat when I get 
access to a windows machine
so I can test it....


>   Instead, I commented out line 85 of src/jsint/Evaluator.java:
>
>    //error.println("\n\n"+Version.VERSION+"\n\n");
>
> Now, src/build/bootstrap works.
>
> Personally, I'm happy enough with just a "> " prompt when I start up 
> JScheme.  But maybe there should be a (help) or (jscheme-version) 
> users could type if they're going through Emacs buffers and can't 
> remember which ones are running JScheme / mzscheme / MIT/GNU Scheme / 
> Kawa / Guile / SISC / ...

You can always type jsint.Version.VERSION$
but we probably do want a simpler way to access the version, e.g. 
(jscheme-version)
(or how about    (JScheme.version) , i.e. a static method of 
jscheme.JScheme.java

a (help) or (help x) might be nice too....
(apropos x) goes a long way toward helping.... as does (describe x)

---Tim---
>
> Geoffrey
> --
> Geoffrey S. Knauth | http://knauth.org/gsk
>
>
>
> -------------------------------------------------------
> This SF.Net email sponsored by Black Hat Briefings & Training.
> Attend Black Hat Briefings & Training, Las Vegas July 24-29 - digital 
> self defense, top technical experts, no vendor pitches, unmatched 
> networking opportunities. Visit www.blackhat.com
> _______________________________________________
> Jscheme-devel mailing list
> Jsc...@li...
> https://lists.sourceforge.net/lists/listinfo/jscheme-devel

Re: [Jscheme-devel] src/jsint/Evaluator.java:85 cannot find Version class

[Geoffrey K. <ge...@kn...>]

On Jun 24, 2004, at 03:44, I wrote:
> For now I'll just make my own Version.java using what I see in 
> (make-version).

I take that back.  It seems Version.java was really supposed to go 
away, and (make-version) is out of date.  Instead, I commented out line 
85 of src/jsint/Evaluator.java:

    //error.println("\n\n"+Version.VERSION+"\n\n");

Now, src/build/bootstrap works.

Personally, I'm happy enough with just a "> " prompt when I start up 
JScheme.  But maybe there should be a (help) or (jscheme-version) users 
could type if they're going through Emacs buffers and can't remember 
which ones are running JScheme / mzscheme / MIT/GNU Scheme / Kawa / 
Guile / SISC / ...

Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk

[Jscheme-devel] src/jsint/Evaluator.java:85 cannot find Version class

[Geoffrey K. <ge...@kn...>]

[ following CVS checkout ]

--------------------------------
$ cd jscheme
$ src/build/bootstrap
src/jsint/Evaluator.java:85: cannot resolve symbol
symbol  : variable Version
location: class jsint.Evaluator
     error.println("\n\n"+Version.VERSION+"\n\n");
                          ^
1 error
Exception in thread "main" java.lang.NoClassDefFoundError: 
jscheme/SchemeEvaluator
	at jscheme.REPL.main(REPL.java:138)
--------------------------------

I see (make-version) in jscheme-bootstrap.scm, but it looks as though 
it hasn't been called before Evaluator.java needs it.  For now I'll 
just make my own Version.java using what I see in (make-version).

Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk

[Jscheme-devel] Scheme Readers, was ... guile's quasi string approach

[Timothy J. H. <tim...@ma...>]

On Jun 23, 2004, at 6:35 PM, Ken Anderson wrote:

> I think the idea that JScheme is an alternative "skin" for java 
> programs is an important one, and your idea of allowing alternative 
> languages to provide java behavior by using JScheme as a back end is a 
> good one.
>
> We could support multiple syntax by allowing a customizable #macro, or 
> as Common Lisp did, allow a customizable (read) so you could change 
> its read-table.

One way to allowing multiple syntaxes is just provide a reader 
(stream->s-expressions+errors)
and then to provide a way of setting the reader for a JScheme instance. 
For example, one might
say

 >  (define js (jscheme.JScheme.))    ;; create a JScheme instance
 >  (.setReader js infix-reader)            ;; set its reader to infix
 >  (.eval js #{ square(x) := x*x }#)      ;; define the square function

 >  (let ((z (.eval js 'square)))               ;; grab that function 
and apply it to 35
       (z 35))

 >  (.setReader js norvig-reader)            ;; now set the reader to 
norvig-reader
 >  (.eval js #{ define(sq(x),*(x,x)) }# )   ;; define square with this 
syntax
                                                                   ;; 
note that define is a macro which operates after reading...
 >   ((.eval js 'sq) 35)                                 ;; apply it,

 >   (.REPL js)                                           ;; now start a 
norvig-reader REPL ...
    > define(fib(n),if(<(n,3),1,+(fib(-(n,1)),fib(-(n,2)))))
    > fib(10)
    > exit()
 >


>
> After TCL/TK came out Stallman i believe criticized it for being a bad 
> scripting language, because string hacking was everywhere, especially 
> in the inner loop.
> He said something like "translate TCL into Scheme and win big".  I 
> believe the outcome was that the people who tried to translate TCL 
> into Scheme had to make semantic changes to TCL for things to work 
> out, much as PLT Scheme needs changes to Python for their analysis to 
> be userful.
This might not be a bad thing if one is thinking of creating a new 
language that combines
Python syntax and Scheme semantics....

>  I'm sure that the amount of effort provided from the Scheme side to 
> the other language was substantial.
>
> Eventually, someone wrote a TCL "compiler" that ran at about as well 
> as the Scheme implementation.  Which just proves that bad ideas will 
> continue to win as long as someone is willing to put more effort into 
> them.
>
> While you can translate the code into Scheme, its harder to translate 
> the error messages.
True...

>   Why am i getting a JScheme and Java backtracke when i'm talking to 
> prolog?
I was actually thinking about defining new readers for Scheme
so the syntax would be infix (or operator/precedence/associativity) but 
the semantics would be Scheme.
The idea about compiling language L into Scheme is interesting too but 
much more complex, I was
thinking more about low hanging fruit.  I think Peter Norvig once said 
that he provided a LISP reader
that converted  F(A,B,...,Z)  to (F A B ... Z) and won some converts to 
Scheme on that simple change
alone, e.g.
   define( fib(n),   if(  <(n,3),  1,   +(  fib(-(n,1)),  fib(-(n,2)))))

This makes it look like a non-LISP syntax, but its really just a slight 
twist on s-expressions.


>   We can hide this to some extent, but to provide "good" language 
> specific runtime error message may be hard.  Of course, the JScheme 
> attitude should be, that that doesn't make it worth trying!  JScheme - 
> you can get a lot more than you might expect from a little language.

Moreover, it may be worthwhile creating a Python-like language which 
looks like Python but
compiles to Scheme, or a Java-like language that compiles to Scheme...

Another alternative is just to explore clean interfaces between Java 
versions of Python, Ruby, Prolog,
etc.  and use JScheme to be the glue that holds them all together. 
(When I get some time I want
to create a SISC procedure for invoking JScheme eval, this would allow 
users of SISC to
have full access to JScheme and the javadot notation. Likewise, 
providing links from Jython, JRuby,
and Java versions of Prolog would be interesting too.....)

---Tim---


>
>
>
> -------------------------------------------------------
> This SF.Net email sponsored by Black Hat Briefings & Training.
> Attend Black Hat Briefings & Training, Las Vegas July 24-29 -
> digital self defense, top technical experts, no vendor pitches,
> unmatched networking opportunities. Visit www.blackhat.com
> _______________________________________________
> Jscheme-user mailing list
> Jsc...@li...
> https://lists.sourceforge.net/lists/listinfo/jscheme-user

[Jscheme-devel] (import "yada.yada.Yada")

[Geoffrey K. <ge...@kn...>]

I think I should get an error when an import is invalid or fails:

 > (import "java.io.File")
#t   ; OK
 > (import "yada.yada.Yada")
#t   ; Huh?

Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk

Re: [Jscheme-devel] (eq? 128 128) => #f

[Ken A. <kan...@bb...>]

object equality is one of those things that seem like it should be simple, but it gets pretty complicated after you think about it, especially if you thinking about it in terms of parallel or distributed computation.  Here's a nice paper by Henry Baker about equality:
http://home.pipeline.com/~hbaker1/ObjectIdentity.html

What it means to copy an object is another example.

Scheme has 4 equality procedures eq? eqv? = and equal? while Java has only = and .equals.  If you only want to use only 2 procedures, use eqv? and equal?.

Here's what  
http://www.schemers.org/Documents/Standards/R5RS/HTML/r5rs-Z-H-9.html#%_idx_216 says about eqv?:

Briefly, it returns #t if obj1 and obj2 should normally be regarded as the same object. This relation is left slightly open to interpretation, but the following partial specification of eqv? holds for all implementations of Scheme. 

There's a lot of wiggle room to allow different Schemes to have different implementation choices.  For example, in a C based implementation of Scheme low order bits (typically) are used for type information so objects like small integers, small floats, characters ... can be implemented as one word "objects", rather than a pointer to memory.

In JScheme, every object is a Java object, but some objects like ascii characters and small integers are cached so eq? may seem to work when you really should be using eqv?.

One difference between JScheme and other Schemes is that literal strings are eq? following Java semantics.

Also, simple tests can full you, so for example (eq? 3 3) may return #t but
(eq? 123456789 123456789) may return #f.

One nice thing you can do in Common Lisp is create a hashtable and give it the equality operator to use.  Java only provides .equals HashTables and HashSets.
k
At 11:02 AM 6/17/2004 -0400, Geoffrey Knauth wrote:
>Never mind.  I should have tried eqv? too.  For the record:
>
>JScheme:
>(eq? 128 128)  => #f            ;; 2^8
>(eq? 127 127)  => #t
>(eqv? 128 128)  => #t
>
>PLT Scheme v207:
>(eq? 1073741824 1073741824)  => #f      ;; 2^30
>(eq? 1073741823 1073741823)  => #t
>(eqv? 1073741824 1073741824)  => #t
>
>Geoffrey
>--
>Geoffrey S. Knauth | http://knauth.org/gsk
>
>On Jun 17, 2004, at 06:50, Geoffrey Knauth wrote:
>
>>I ran into this while calculating azimuths.
>>
>>> (eq? 180 180)
>>#f
>>> (= 180 180)
>>#t
>>> (eq? 0 0)
>>#t
>>> (eq? 1 1)
>>#t
>>> (eq? 2 2)
>>#t
>>> (eq? 3 3)
>>#t
>>> (eq? 180 180)
>>#f
>>
>>After poking around, I found this:
>>
>>> (eq? 127 127)
>>#t
>>> (eq? 128 128)
>>#f
>>> (eq? -127 -127)
>>#t
>>> (eq? -128 -128)
>>#f
>>
>>For now I'll use = instead of eq?.  I was just wondering if this was documented.
>>
>>Geoffrey
>>--
>>Geoffrey S. Knauth | http://knauth.org/gsk
>>
>>
>>
>>-------------------------------------------------------
>>This SF.Net email is sponsored by The 2004 JavaOne(SM) Conference
>>Learn from the experts at JavaOne(SM), Sun's Worldwide Java Developer
>>Conference, June 28 - July 1 at the Moscone Center in San Francisco, CA
>>REGISTER AND SAVE! http://java.sun.com/javaone/sf Priority Code NWMGYKND
>>_______________________________________________
>>Jscheme-devel mailing list
>>Jsc...@li...
>>https://lists.sourceforge.net/lists/listinfo/jscheme-devel
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by The 2004 JavaOne(SM) Conference
>Learn from the experts at JavaOne(SM), Sun's Worldwide Java Developer
>Conference, June 28 - July 1 at the Moscone Center in San Francisco, CA
>REGISTER AND SAVE! http://java.sun.com/javaone/sf Priority Code NWMGYKND
>_______________________________________________
>Jscheme-devel mailing list
>Jsc...@li...
>https://lists.sourceforge.net/lists/listinfo/jscheme-devel

[Jscheme-devel] Version 7.0 and Servlets

[Timothy J. H. <ti...@cs...>]

I ran into a little problem using the new jscheme.jar for servlets. It 
seems that Jetty
creates several threads for handling the servlets and each thread now 
creates its own
interpreter which has its own environment!

Before,  I could just do a define in one servlet and all later servlets 
would have access to the
definition. I may be able to switch back to that model (by using the 
jscheme.JS class instead
of jscheme.JScheme), but its an interesting example of the kind of bugs 
that can arise when
adding new features....

Another fix is to explicitly import any modules that you need when you 
generate a webpage.
Since modules are cached, this won't slow down the servlet engine much 
and it is cleaner
than my old adhoc approach...

 From a development point of view though, this will require the addition 
of a "reload-module"
procedure that disregards the caching when a module is changed.....

---Tim---

[Jscheme-devel] Re: [Jscheme-user] Hash-Quasi-Strings and Fixed Modulo Primitive

[Timothy J. H. <ti...@cs...>]

On Jun 17, 2004, at 1:07 PM, Ken Anderson wrote:

> Just to be clear my proposal was to use #{ ... }# and #[ ...]#. (ie # 
> is outside).
>
> Well, there's a tradition in common lisp, at least that we'd be 
> against if we use
> #{ ... #} and #[ ... #].
>
> In CL you have:
> #<... ># to indicate the print representation of a thing that can be 
> read by (read).
> #| ... |#  for an extended comment.  I think PLT Scheme has this too.

I see. The ide is that the initial token is HASH-SYMBOL and the close 
token is
just the reverse.....

Since there are precedences in Common Lisp and PLT scheme, not to 
mention the
/* ... */ for Java or the case ... esac of Ada, I think it makes sense.

I'll work on changing the syntax, before anyone has become too heavily 
invested in it....

I've already used this syntax in my class (but I noticed that I 
frequently wrote ]# instead of #] ...
maybe its a sign....).

---Tim---

>
> Though vectors are #(1 2 3) not #(1 2 3)#.
>
> While #{ is a macro character, #} really isn't, and i don't think its 
> right think of it as one.
>
> Johnathan, or Rusty do you have an opinion on this?
>
> At 06:29 PM 6/16/2004 -0400, Timothy John Hickey wrote:
>> I guess my feeling was that it might get confusing for people to 
>> remember
>> whether we were using #{... }#  or {#....#}
>> and if we escaped with #[...]# or [#...#]
>> So having a simple HASH-SYMBOL syntax for each delimiter makes it 
>> somewhat easier
>> to remember...
>

[Jscheme-devel] Re: [Jscheme-user] Hash-Quasi-Strings and Fixed Modulo Primitive

[Ken A. <kan...@bb...>]

Just to be clear my proposal was to use #{ ... }# and #[ ...]#. (ie # is outside).

Well, there's a tradition in common lisp, at least that we'd be against if we use
#{ ... #} and #[ ... #].

In CL you have:
#<... ># to indicate the print representation of a thing that can be read by (read).
#| ... |#  for an extended comment.  I think PLT Scheme has this too.

Though vectors are #(1 2 3) not #(1 2 3)#.

While #{ is a macro character, #} really isn't, and i don't think its right think of it as one.

Johnathan, or Rusty do you have an opinion on this?

At 06:29 PM 6/16/2004 -0400, Timothy John Hickey wrote:
>I guess my feeling was that it might get confusing for people to remember
>whether we were using #{... }#  or {#....#}
>and if we escaped with #[...]# or [#...#]
>So having a simple HASH-SYMBOL syntax for each delimiter makes it somewhat easier
>to remember...

Re: [Jscheme-devel] (eq? 128 128) => #f

[Geoffrey K. <ge...@kn...>]

Never mind.  I should have tried eqv? too.  For the record:

JScheme:
(eq? 128 128)  => #f		;; 2^8
(eq? 127 127)  => #t
(eqv? 128 128)  => #t

PLT Scheme v207:
(eq? 1073741824 1073741824)  => #f	;; 2^30
(eq? 1073741823 1073741823)  => #t
(eqv? 1073741824 1073741824)  => #t

Geoffrey
--
Geoffrey S. Knauth | http://knauth.org/gsk

On Jun 17, 2004, at 06:50, Geoffrey Knauth wrote:

> I ran into this while calculating azimuths.
>
> > (eq? 180 180)
> #f
> > (= 180 180)
> #t
> > (eq? 0 0)
> #t
> > (eq? 1 1)
> #t
> > (eq? 2 2)
> #t
> > (eq? 3 3)
> #t
> > (eq? 180 180)
> #f
>
> After poking around, I found this:
>
> > (eq? 127 127)
> #t
> > (eq? 128 128)
> #f
> > (eq? -127 -127)
> #t
> > (eq? -128 -128)
> #f
>
> For now I'll use = instead of eq?.  I was just wondering if this was 
> documented.
>
> Geoffrey
> --
> Geoffrey S. Knauth | http://knauth.org/gsk
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by The 2004 JavaOne(SM) Conference
> Learn from the experts at JavaOne(SM), Sun's Worldwide Java Developer
> Conference, June 28 - July 1 at the Moscone Center in San Francisco, CA
> REGISTER AND SAVE! http://java.sun.com/javaone/sf Priority Code 
> NWMGYKND
> _______________________________________________
> Jscheme-devel mailing list
> Jsc...@li...
> https://lists.sourceforge.net/lists/listinfo/jscheme-devel
>