Path: ...!news-out.netnews.com!postmaster.netnews.com!us7.netnews.com!not-for-mail
X-Trace: DXC=g7EB\AmPd3cPHaK9SYdC3bHWonT5<]0TmQ;nb^V>PUff5[gZBW6J?LlTD;10@6=i=nZ[0LWn0`0Eb>G9_h6]J@kjWjB\<Me8\Ag@GM\IYO6WZd>Oa:X5b?f8h
X-Complaints-To: support@blocknews.net
Date: Sun, 16 Jun 2024 14:30:11 -0400
MIME-Version: 1.0
User-Agent: Betterbird (Windows)
From: DFS <nospam@dfs.com>
Subject: Re: C Non-Programming Non-Challenge
Newsgroups: comp.os.linux.advocacy
References: <17d933d9ffc8ee40$16989$3694546$802601b3@news.usenetexpress.com>
 <17d943beaefb3aa2$16097$2041738$802601b3@news.usenetexpress.com>
 <rPadnWt9mtOO9PP7nZ2dnZfqn_qdnZ2d@supernews.com>
 <ld7lfuF433iU2@mid.individual.net>
Content-Language: en-US
In-Reply-To: <ld7lfuF433iU2@mid.individual.net>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Lines: 148
Message-ID: <666f2f33$0$3102233$882e4bbb@reader.netnews.com>
NNTP-Posting-Host: 127.0.0.1
X-Trace: 1718562611 reader.netnews.com 3102233 127.0.0.1:49193
Bytes: 5181

On 6/16/2024 4:19 AM, vallor wrote:
>> On Jun 15, 2024 at 3:20:19 PM EDT, "Farley Flud" <ff@linux.rocks> wrote:
>>
>>> On Sat, 15 Jun 2024 14:29:04 +0000, Farley Flud wrote:
>>>
>>>
>>>> Write a C program to compute the subfactorial of an integer N.
>>>>
>>>>
>>> Nobody got it (as predicted).
>>
>> Because no one bothered.
> 
> I actually took the opportunity to install the GMP documentation
> on my system, look up what a "subfactorial" is, and do
> an implementation that works for !N where N is 0 through 50.

Where's your code?

Here's an approximation of !n, using round(n!/e)
-----------------------------------------
#include <stdio.h>
#include <stdlib.h>
int main(int argc, char *argv[]){
	int n = atoi(argv[1]);
	for(int i=1;i<=n;i++){
		double f=1.0;
		for(int n=1;n<=i;n++){
			f*=n;
		}
		f/=2.71828;
		printf("%d. %.0f\n",i,(f));
	}
}	
-----------------------------------------
$ ./prog n

Without using a bignum library, it craps out at 171 and prints 'inf' (so 
does python)


1. 0
2. 1
3. 2
4. 9
5. 44
6. 265
7. 1854
8. 14833
9. 133496
10. 1334962
11. 14684580
12. 176214959
13. 2290794473
14. 32071122622
15. 481066839325
16. 7697069429198
17. 130850180296364
18. 2355303245334550
19. 44750761661356448
20. 895015233227128960
21. 18795319897769705472
22. 413497037750933585920
23. 9510431868271472934912
24. 228250364838515316883456
25. 5706259120962883593175040
....
50. 11188727136907667116392988817975029667335257249097046505283387392
....
170. 
2669855796793558290010001570575275101705027811519977073818130755407973315261224007251979700779512228624245141462623296456740188588977348681648565542706396927622340485993465997718765185927438606647566078414539613618036013433690401156103581768582773608415498699128984156662958003889715092930354450116728848384
171. inf


> Furled Fart lost the opportunity to show his prowess with
> the GMP library by not demonstrating use of gmp_printf().

Note he also lied "we ignore N = 1, 2 as they are trivial"

He ignored them because he didn't know how to handle those cases. 
That's the kind of devious, evasive wanker he is.

Notice he said "you missed a definite error in the code that I 
deliberately placed to catch any sharp eyes, but no catch."

This is another devious lie he uses to try and excuse mistakes in his 
"perfect code".



>> This is all trivial stuff.  These math "problems" have already been
>> solved multiple times.  That you think this is a "programming challenge"
>> just shows how utterly clueless you are.
> 
> It's more of a "let's learn libgmp" challenge, which is a library
> that he already has familiarity with.  I don't think he learned anything.
> And it's not that difficult a library to use.
> 
>> Hint: people were doing this 60 years ago on mainframes using Fortran.
>> That you have only now discovered this proves AGAIN how utterly clueless
>> you are.
>>    
>>    
>> What's next? Compute Pi to a million decimal places? That also has
>> already been done many times. Sounds like another great "programming
>> challenge" for your feeble brain.
> 
> $ time ./try_mpz
> 0:1
> 1:0
> 2:1
> 3:2
> 4:9
> 5:44
> [...]
> 49:223774392215649610092605161415154686480227084177314431854406736
> 50:11188719610782480504630258070757734324011354208865721592720336801
>
> real	0m0.002s
> user	0m0.001s
> sys	0m0.001s
> 
> I also was interested in the recursive version.  Here's the
> run with N=50 with long long int's, which gives the wrong answer
> due to overflow:
> 
> $ time ./try2
> 7531320031745615777
> 
> real	0m35.541s
> user	0m35.530s
> sys	0m0.001s
> 
> So: don't bother with the recursive version, except as
> a pedantic exercize.
> 
> *Do* use GMP if you ever need to work with big numbers...
> maybe numpy in Python can do the same thing, I wouldn't
> know about that.
> 
> $ ll /usr/share/doc/gmp-doc/gmp.pdf.gz
> -rw-r--r-- 1 root root 812615 Nov 16  2020 /usr/share/doc/gmp-doc/
> gmp.pdf.gz
> 
> $ grep FLAGS Makefile
> CFLAGS=-Wall -O3 -g -Wpedantic
>