Should create a strong password:
variable _index [pid];
proc password {length} {
global _index;
variable _set {shW4UMe832TpaSylIdfzxbrNki9A7Q5PmZ0XDuYVg1KEGwLcJtjRCF6OqovBHn};
variable _password {};
for {set _char 1} {$_char <= $length} {incr _char} {
incr _index [lindex [time {
while {$_index >= [string length $_set]} {
incr _index -[string length $_set];
}
# Only usefull to increase entropy
for {set _count $_index} {$_count <= [string length $_set]+[string index [clock microseconds] end]} {incr _count} {
# Nothing to do
}
set _password $_password[string index $_set $_index];
}] 0]
}
return $_password;
}
Call password length inside the script.
AMG: My first instinct was to optimize this code since it is written in a seemingly wasteful way. However, there is a trick, a method to the madness. The time command is used as a random number generator, used to advance the index into the character set. So inefficient programming with somewhat unpredictable runtime is exactly the programmer's intent.
bll 2016-4-15:
See [L1 ]:
The output from "ent" for 80000 characters, password 8 called 10000 times. The serial correlation coefficient is quite high.
Entropy = 4.787082 bits per byte. Optimum compression would reduce the size of this 80000 byte file by 40 percent. Chi square distribution for 80000 samples is 841393.86, and randomly would exceed this value 0.01 percent of the times. Arithmetic mean value of data bytes is 93.7709 (127.5 = random). Monte Carlo value for Pi is 4.000000000 (error 27.32 percent). Serial correlation coefficient is -0.204070 (totally uncorrelated = 0.0).
paskali 2016-04-15:
Thanks. I am crazy.
paskali - 2016-04-16 13:38:58
Update, now the previous test should be something like this:
Entropy = 5.953689 bits per byte. Optimum compression would reduce the size of this 80000 byte file by 25 percent. Chi square distribution for 80000 samples is 250555.01, and randomly would exceed this value less than 0.01 percent of the times. Arithmetic mean value of data bytes is 86.9527 (127.5 = random). Monte Carlo value for Pi is 4.000000000 (error 27.32 percent). Serial correlation coefficient is 0.000435 (totally uncorrelated = 0.0).
bll 2016-4-16
The latest version:
See [L2 ]:
The output from "ent" for 80000 characters, password 8 called 10000 times. The serial correlation coefficient is very high. As seen from the character occurrence table, this is not a strong generator.
Entropy = 3.753542 bits per byte. Optimum compression would reduce the size of this 80000 byte file by 53 percent. Chi square distribution for 80000 samples is 1668541.12, and randomly would exceed this value 0.01 percent of the times. Arithmetic mean value of data bytes is 101.3792 (127.5 = random). Monte Carlo value for Pi is 4.000000000 (error 27.32 percent). Serial correlation coefficient is -0.048597 (totally uncorrelated = 0.0).
The character occurence table:
Value Char Occurrences Fraction 48 0 64 0.000800 49 1 31 0.000387 50 2 12 0.000150 51 3 17 0.000212 52 4 11 0.000138 53 5 122 0.001525 54 6 11 0.000138 55 7 183 0.002288 56 8 18 0.000225 57 9 1063 0.013288 65 A 277 0.003462 66 B 12 0.000150 67 C 17 0.000212 68 D 64 0.000800 69 E 25 0.000313 70 F 17 0.000212 71 G 15 0.000188 72 H 10 0.000125 73 I 6791 0.084888 74 J 13 0.000162 75 K 24 0.000300 76 L 17 0.000212 77 M 11 0.000138 78 N 7556 0.094450 79 O 14 0.000175 80 P 111 0.001388 81 Q 152 0.001900 82 R 7 0.000087 83 S 309 0.003863 84 T 17 0.000212 85 U 10 0.000125 86 V 41 0.000513 87 W 12 0.000150 88 X 62 0.000775 89 Y 52 0.000650 90 Z 98 0.001225 97 a 120 0.001500 98 b 7842 0.098025 99 c 13 0.000162 100 d 7534 0.094175 101 e 15 0.000188 102 f 7720 0.096500 103 g 21 0.000262 104 h 15 0.000188 105 i 3328 0.041600 106 j 21 0.000262 107 k 7020 0.087750 108 l 4747 0.059338 109 m 102 0.001275 110 n 8 0.000100 111 o 13 0.000162 112 p 15 0.000188 113 q 11 0.000138 114 r 7738 0.096725 115 s 15 0.000188 116 t 20 0.000250 117 u 46 0.000575 118 v 12 0.000150 119 w 19 0.000237 120 x 7718 0.096475 121 y 817 0.010212 122 z 7804 0.097550 Total: 80000 1.000000