newlispfanclub.alh.net

Has anyone tried to compile newlisp statically for use on Linux or other Unix? It may work on BSD and Mac OSX, I haven't tried. However, the problem is always the resolver. If you use any network functions in the program, you simply cannot get a static binary. What is a static binary? A static binar...

In my /dev/urandom function defined over in this thread http://www.newlispfanclub.alh.net/forum/viewtopic.php?f=16&t=4864 it would be useful to have a destructor. Purpose of destructor would be to do (close fd) to close the file descriptor. But then I thought about it; how many "random number genera...

I developed this function for some population simulations I'm running. Newlisp has a nice built-in random number generator. But I wanted a bit more randomness than that provides. So I wrote this function to access the urandom device on Linux, which gives you lots of pretty good random numbers, fast....

We have a flood of spammers bumping topics. But anyway... wow. artfulcode is gone. :(

Thanks HPW!

HPW, thank you. I never used newlisp-tk back when it was maintained. Can you explain how it works? Is it an exe that combines tcl/tk and newlisp into one binary, and then tcl/tk shows up as a "tk" or "tcl" function that you send strings to? What is involved in updating it? I'd like to use it on Linu...

Lutz, what happened to newlisp-tk? Was it similar to tkinter in Python? Did the wrapper code stop working?

jopython wrote:Are you expecting that newlisp bundles a SSL/TLS library?

That would be nice. libressl and libtls are good candidates for inclusion.

We already depend on libpcre I think.

First, does it work without your modifications.

Second, show how you are calling the function with your modifications. And show how this differs from the call to the unmodified function.

It shouldn't do anything, all you are doing is defining a couple functions.

Show us what you tried, please.

I use the curl wrapper, works well for me.

I made a TLS wrapper so you could open a TLS connection, but it wasn't integrated into get-url. My wrapper used libtls from the OpenBSD project.

Maybe this will help you. Don't think you need nL at all. Here is some code that does something very similar, and should answer your questions: http://reactor-core.org/baccarat.html If you want to use nL to store the clicks in a database or something, then use AJAX http://www.newlisp.org/index.cgi?p...

If someone codes one up, I'll use it.

Although I'm more interested in FastCGI support in newlisp, then I can use newlisp to power a website using OpenBSD's small/secure new webserver. As well as Apache/nginx and all the other mainstream web servers.

That code looks correct. What is the error you are experiencing?

Thank you Ralph, that explains it.

Update:

Although, that is a "surprise". I expect dolist to return a list, then suddenly it doesn't... I'd expect nil in case of an error, but the empty list makes more sense as a return value.

ssqq, yes, thank you. I did similar little tests to see that extend should work the way I am using it. But in the code I posted, it doesn't. Any ideas? Are the debug messages sending me barking up the wrong tree?

Also thanks for reminding me of the "count" function. I tried to reimplement it in an ugly way using map. Glad it is builtin. I thought "there should be something to do this built in to the language." I should have spent 20 minutes going down the list of functions. The newLisp API is indeed rich and...

Hi Ted! Actually, there is a way to make your program even faster -- a lot faster. The idea is that there is a way to determine the number of divisors of a (natural) number which is a lot cheaper than computing the unique sub-groupings of its prime factors. Here is a routine that does just that. .....

Thank you Lutz. Mystery solved. Also, you may want to update the Coding Patterns document to use the version of "combinations" function in my last post; it is quite a lot more efficient. Saves one line of code too. (extend) is a real bonus function.

I changed the "combinations" function to use the (extend foo bar) idiom instead of (setq foo (append foo bar)) The code is actually running faster, and using less memory too. Instead of 18% of memory, it is using 3%. This is still a lot more than I expect, so I am curious what is going on, if it is ...

“extend” would treat first argument as place and Increase by degress,you'd better use cons or other ... Do you mind showing an example of what you mean? I thought (extend foo bar) was supposed to be more (or as) efficient as (setq foo (append foo bar)) Newlisp doesn't have cons cells. Also, if you ...

Sorry, that function was in my init.lsp file. I have a lot of little utility functions there.
I've added it into the original post now. Should work.

-nan (NaN) should not return true with number? > (set 'nan (sqrt -1)) > (number? nan) true ;; should return nil Also advise with inf. I think all NaN and Inf are "error status" that may not throw error immediatly. So they isn't value. Reliably getting nan and inf from the C compiler is not pleasant...

I wrote some code to generate the list of antiprimes. Had fun with it. Sped it up by quite a large factor by using memoization. But it blew up my memory usage beyond what I was expecting. So I dropped back to an algorithm that I expected to not take up much memory at all, less than a megabyte. But 2...