[Bugsy]

This is basically what Natsheh suggested in his first post:

PHP Code:


#include <amxmodx>
#include <amxmisc>

const MaxNumber = 30;
const ExcludeNumbers = ( ( 1 << 13 ) | ( 1 << 22 ) );

new g_RandomNumbers[ MaxNumber ];

public plugin_init()
{
    LoadRandomNumbers();
    
    //Access your random number with this: g_RandomNumbers[ random( MaxNumber ) ]
    server_print( "Random=%d" , g_RandomNumbers[ random( MaxNumber ) ] );
}

public LoadRandomNumbers()
{
    new iNumber , iIndex;
    
    while ( ( iIndex < MaxNumber ) && ( iNumber <= MaxNumber ) )
    {
        if ( !( ExcludeNumbers & ( 1 << iNumber ) ) )
        {
            g_RandomNumbers[ iIndex++ ] = iNumber;
        }
        
        iNumber++;
    }
} 


[Natsheh]

Originally Posted by Bugsy This is basically what Natsheh suggested in his first post: PHP Code: #include <amxmodx> #include <amxmisc> const MaxNumber = 30; const ExcludeNumbers = ( ( 1 << 13 ) | ( 1 << 22 ) ); new g_RandomNumbers[ MaxNumber ]; public plugin_init() {     new size = LoadRandomNumbers();          //Access your random number with this: g_RandomNumbers[ random( size) ]     server_print( "Random=%d" , g_RandomNumbers[ random( size ) ] ); } public LoadRandomNumbers() {     new iNumber , iIndex, size;          while ( ( iIndex < MaxNumber ) && ( iNumber <= MaxNumber ) )     {         if ( !( ExcludeNumbers & ( 1 << iNumber ) ) )         {             g_RandomNumbers[ iIndex++ ] = iNumber;                         Size++;         }                 iNumber++;     }         return size; }

Nope that not quite right i edited the post

[Bugsy]

Originally Posted by Natsheh Nope that not quite right i edited the post

That's great, lol. But the code works fine.

[fysiks]

Originally Posted by Black Rose The chance of that happening in that span with so few exceptions is so small that it's not even worth caring about.

You are correct that the chance of an infinite loop happening is extremely small. However, that's not the problem when it comes to this type of thing. The problem is that the number of loops required is ALWAYS unknown (it could be 1 and it could be 500 or worse). This is called non-deterministic execution. You should always strive for deterministic execution especially because most code executed in plugins is blocking.

All of this was discussed thoroughly in the thread noted by PRoSToTeM@.

[Black Rose]

Originally Posted by fysiks You are correct that the chance of an infinite loop happening is extremely small. However, that's not the problem when it comes to this type of thing. The problem is that the number of loops required is ALWAYS unknown (it could be 1 and it could be 500 or worse). This is called non-deterministic execution. You should always strive for deterministic execution especially because most code executed in plugins is blocking. All of this was discussed thoroughly in the thread noted by PRoSToTeM@.

You don't discuss fysiks, you tell people.

It could be 500, but it won't be. And even if it would be 500, though the chance is 0.0000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000 00009% (yes, that is the actual chance), you wouldn't even notice it because 500 cycles of such a small code is done in less than one millisecond.

I get it, it's good to avoid the unknown, but only when it's reasonable to do so.

I assume you agree with my previous example. It would be stupid to cache 1.45TiB memory and do one guaranteed cycle. But the same logic still applies.
Where do you draw the line?
How much is too much to buffer to enter the unknown?
Or is it a question of probability instead of memory consumption because -9*10⁵⁸⁹ to 1 didn't seem enough.

Another example. Let's say you want 10 random points in a circle from a point in a 2 dimensional world. How would you go about that with your fancy lookup table?

The thing is, you often tell people what to do but you never give any other reason except "it's the right way".

[fysiks]

Originally Posted by Black Rose I get it, it's good to avoid the unknown, but only when it's reasonable to do so.

And in this case, it is 100% reasonable to do it the deterministic way. Obviously, the person that writes their own code as final say as to what is actually put in their code. So, they get to evaluate the various methods proposed and choose the one that the want.

[PRoSToTeM@]

Originally Posted by Black Rose Randomizing a temporary filename with 6 characters, assuming we use all lowercase, uppercase and numeric characters and want to avoid an infinite loop we would end up with 26+26+10 different characters ^6, therefor 56 800 235 584 different options. Saving those in an array of 7 cells each since you need to null-terminate that's 1.45Terabyte of memory.

This method should work good here, it requires strong order (maybe weak order can work too), which strings have.

[indraraj striker]

lol. thanks everyone
#solved

[Black Rose]

I know it proves nothing, just for fun.

#include <amxmodx>
#include <rose>

public plugin_init() {
    register_plugin("Test Plugin 1", "1.0", "[ --{-@ ]");
    return;
    new szTimer[16];

    new record, totalcycles, totalloops, hTimer = TimerStart();

    while ( record < 500 ) {
        new i, a;
        totalcycles++;
        do {
            i++;
            totalloops++;
            a = random_num(0, 30);
        }
        while (a == 13 || a == 22);

        if ( i > record ) {
            TimerFormat(TimerMid(hTimer), szTimer, charsmax(szTimer), 1, true);
            server_print("%s: New record: %d loops before exiting. (Total cycles: %d, total loops: %d, avg loops: %.2f)", szTimer, i, totalcycles, totalloops, 1.0 * totalloops / totalcycles);
            record = i;
        }
    }
}

/*
00:00:00:00.000: New record: 2 loops before exiting. (Total cycles: 1, total loops: 2, avg loops: 2.00)
00:00:00:00.000: New record: 3 loops before exiting. (Total cycles: 352, total loops: 375, avg loops: 1.06)
00:00:00:00.001: New record: 4 loops before exiting. (Total cycles: 2017, totalloops: 2183, avg loops: 1.08)
00:00:00:00.005: New record: 5 loops before exiting. (Total cycles: 11173, total loops: 11943, avg loops: 1.06)
00:00:00:00.306: New record: 6 loops before exiting. (Total cycles: 1247442, total loops: 1333647, avg loops: 1.06)
00:00:00:03.651: New record: 7 loops before exiting. (Total cycles: 15484486, total loops: 16552027, avg loops: 1.06)
00:00:01:24.609: New record: 8 loops before exiting. (Total cycles: 360494522, total loops: 385358373, avg loops: 1.06)
00:00:20:56.957: New record: 9 loops before exiting. (Total cycles: 1055510066,total loops: 1424513328, avg loops: 1.34)
00:00:21:59.124: New record: 10 loops before exiting. (Total cycles: outside range, total loops: outside range, avg loops: N/A)
03:19:25:xx.xxx: I was only supposed to run it overnight but forgot it for almost 4 days.
 */

[PRoSToTeM@]

Now try to use custom random function that always returns 13. xD