[plt-scheme] conditional module inclusion and compilation

Ah -- (system-type) is nice, and I'm not at all worried about the cost  
of a few calls to void.

But I still don't see how to make this work because I have to have  
(require plot) in the file to make the plots work in macosx, (require  
plot) breaks under Linux ("dynamic-require: unknown module: '#%mred- 
kernel; need to run in mred instead of mzscheme"), and I can't wrap  
(require plot) in a conditional ("require: not at module level or top  
level in: (require plot)"). While I could comment out and uncomment  
the (require plot) I'd have a similar problem with the calls to the  
functions in the plot module, even if they are wrapped in conditions  
that prevent them from being executed in the wrong (system-type).

Just saw that Eli sent some additional info...

  -Lee

On Sep 2, 2009, at 5:52 PM, Matthias Felleisen wrote:

>
> Would
>
>> Welcome to MzScheme v4.2.1.7 [3m], Copyright (c) 2004-2009 PLT  
>> Scheme Inc.
>> > (system-type)
>> macosx
>
> help you get started? If you can tolerate a call to void for plots  
> on the cluster, I'd use an abstraction over the graphics. If you  
> want to see the tracker, pass in draw-progress-report (or whatever  
> you call the graphics function); otherwise pass in void instead.
>
> If the extra call costs too much, I'd write a macro that unfolds in  
> one way on a mac and in a different way on a linux box.
>
> -- Matthias
>
>
>
>
>
>
>
> On Sep 2, 2009, at 5:34 PM, Lee Spector wrote:
>
>>
>> I've developed some code in DrScheme under Mac OS X that I also  
>> want to run sometimes under Linux -- it's computationally intensive  
>> stuff and I'll want to do development runs on my mac but then do  
>> lots of bigger runs on our Linux cluster.
>>
>> I have some graphics in the Mac OS X version that I don't know how  
>> to produce in my command-line/text-based Linux environment, but I  
>> don't need the graphics there anyway. The graphics are just plots  
>> of progress which are useful when I'm developing and watching a  
>> run, but the cluster runs produce text files of data which are good  
>> enough. So to get the code working on the cluster I just removed  
>> "(require plot)" and all of the code that called anything from the  
>> plot module.
>>
>> That worked but I don't want to have to do this kind of surgery on  
>> the code every time I move it to the cluster, which I'll be doing  
>> frequently. Ideally I'd like to be able to write my code in such a  
>> way that it knows where it's running and it includes the plot  
>> module and calls the plot code only if it's running somewhere  
>> appropriate. I do this in Common Lisp by using the #+: and #-:  
>> conditional compilation syntax and/or by looking for things in the  
>> variable *features*, which includes symbols that describe the  
>> environment. Is there anything similar in PLT?
>>
>> I wouldn't mind having to manually change a flag to indicate the  
>> environment, e.g. to change the value in something like:
>>
>> (define include-plots-in-reports #t)
>>
>> But this doesn't actually help for two reasons:
>>
>> - PLT won't let me wrap a call to "require" in a conditional  
>> expression -- it complains that it's not at the top level.
>> - If I comment out the "require" PLT still won't compile a file  
>> containing calls to functions in the missing module, even if those  
>> calls are never executed.
>>
>> So whether my code knows where it's running or whether I have to  
>> tell it manually, I don't have a way of moving it to the Linux  
>> environment without manually editing out the "require" and all of  
>> the code that refers to the plot module.
>>
>> Is there a good way to deal with this kind of thing in PLT that I'm  
>> missing?
>>
>> Thanks,
>>
>> -Lee
>>
>> --
>> Lee Spector, Professor of Computer Science
>> School of Cognitive Science, Hampshire College
>> 893 West Street, Amherst, MA 01002-3359
>> lspector at hampshire.edu, http://hampshire.edu/lspector/
>> Phone: 413-559-5352, Fax: 413-559-5438
>>
>> Check out Genetic Programming and Evolvable Machines:
>> http://www.springer.com/10710 - http://gpemjournal.blogspot.com/
>>
>> _________________________________________________
>> For list-related administrative tasks:
>> http://list.cs.brown.edu/mailman/listinfo/plt-scheme

--
Lee Spector, Professor of Computer Science
School of Cognitive Science, Hampshire College
893 West Street, Amherst, MA 01002-3359
lspector at hampshire.edu, http://hampshire.edu/lspector/
Phone: 413-559-5352, Fax: 413-559-5438

Check out Genetic Programming and Evolvable Machines:
http://www.springer.com/10710 - http://gpemjournal.blogspot.com/