unicode.c

   1 /*
   2  * Routines for generating a page of mangled unicode.
   3  *
   4  * Inspiration:
   5  *  http://www.cl.cam.ac.uk/~mgk25/ucs/examples/quickbrown.txt
   6  *  http://www.columbia.edu/~fdc/utf8/
   7  *  http://www.cl.cam.ac.uk/~mgk25/unicode.html
   8  *  http://www.cl.cam.ac.uk/~mgk25/ucs/examples/UTF-8-test.txt
   9  *  http://stackoverflow.com/questions/1319022/really-good-bad-utf-8-example-test-data
  10  *  http://www.twitter.com/glitchr
  11  *
  12  * Lots more to do here, but this is a start.
  13  */
  14
  15 #include <stdlib.h>
  16 #include <unistd.h>
  17 #include <string.h>
  18 #include "arch.h"
  19 #include "sanitise.h"
  20
  21 void gen_unicode_page(char *page)
  22 {
  23         unsigned int i = 0, j, l;
  24         unsigned int unilen;
  25
  26         const char unicode1[4] = { 0xb8, 0xe0, 0xe0, 0xaa };
  27         const char unicode2[6] = { 0x89, 0xb9, 0xb9, 0xe0, 0xe0, 0x89 };
  28         const char unicode3[2] = { 0x89, 0xb9 };
  29         const char unicode4[18] = { 0xbb, 0xef, 0xd2, 0xa9, 0xd2, 0x88, 0x20, 0x88, 0x88, 0xd2, 0x88, 0xd2, 0xd2, 0x20, 0xd2, 0x88, 0x0a, 0x88 };
  30         const char unicode5[4] = { 0xd9, 0x20, 0xd2, 0x87 };
  31         const char unicode6[4] = { 0xcc, 0x88, 0xd2, 0xbf };
  32         const char unicode7[2] = { 0x0a, 0xbf };
  33         const char *zalgo[] = { "T̫̺̳o̬̜ ì̬͎̲̟nv̖̗̻̣̹̕o͖̗̠̜̤k͍͚̹͖̼e̦̗̪͍̪͍ ̬ͅt̕h̠͙̮͕͓e̱̜̗͙̭ ̥͔̫͙̪͍̣͝ḥi̼̦͈̼v҉̩̟͚̞͎e͈̟̻͙̦̤-m̷̘̝̱í͚̞̦̳n̝̲̯̙̮͞d̴̺̦͕̫ ̗̭̘͎͖r̞͎̜̜͖͎̫͢ep͇r̝̯̝͖͉͎̺e̴s̥e̵̖̳͉͍̩̗n̢͓̪͕̜̰̠̦t̺̞̰i͟n҉̮̦̖̟g̮͍̱̻͍̜̳ ̳c̖̮̙̣̰̠̩h̷̗͍̖͙̭͇͈a̧͎̯̹̲̺̫ó̭̞̜̣̯͕s̶̤̮̩̘.̨̻̪̖͔  ̳̭̦̭̭̦̞́I̠͍̮n͇̹̪̬v̴͖̭̗̖o̸k҉̬̤͓͚̠͍i͜n̛̩̹͉̘̹g͙ ̠̥ͅt̰͖͞h̫̼̪e̟̩̝ ̭̠̲̫͔fe̤͇̝̱e͖̮̠̹̭͖͕l͖̲̘͖̠̪i̢̖͎̮̗̯͓̩n̸̰g̙̱̘̗͚̬ͅ ͍o͍͍̩̮͢f̖͓̦̥ ̘͘c̵̫̱̗͚͓̦h͝a̝͍͍̳̣͖͉o͙̟s̤̞.̙̝̭̣̳̼͟  ̢̻͖͓̬̞̰̦W̮̲̝̼̩̝͖i͖͖͡ͅt̘̯͘h̷̬̖̞̙̰̭̳ ̭̪̕o̥̤̺̝̼̰̯͟ṳ̞̭̤t̨͚̥̗ ̟̺̫̩̤̳̩o̟̰̩̖ͅr̞̘̫̩̼d̡͍̬͎̪̺͚͔e͓͖̝̙r̰͖̲̲̻̠.̺̝̺̟͈  ̣̭T̪̩̼h̥̫̪͔̀e̫̯͜ ̨N̟e҉͔̤zp̮̭͈̟é͉͈ṛ̹̜̺̭͕d̺̪̜͇͓i̞á͕̹̣̻n͉͘ ̗͔̭͡h̲͖̣̺̺i͔̣̖̤͎̯v̠̯̘͖̭̱̯e̡̥͕-m͖̭̣̬̦͈i͖n̞̩͕̟̼̺͜d̘͉ ̯o̷͇̹͕̦f̰̱ ̝͓͉̱̪̪c͈̲̜̺h̘͚a̞͔̭̰̯̗̝o̙͍s͍͇̱͓.̵͕̰͙͈ͅ ̯̞͈̞̱̖Z̯̮̺̤̥̪̕a͏̺̗̼̬̗ḻg͢o̥̱̼.̺̜͇͡ͅ ̴͓͖̭̩͎̗  ̧̪͈̱̹̳͖͙H̵̰̤̰͕̖e̛ ͚͉̗̼̞w̶̩̥͉̮h̩̺̪̩͘ͅọ͎͉̟ ̜̩͔̦̘ͅW̪̫̩̣̲͔̳a͏͔̳͖i͖͜t͓̤̠͓͙s̘̰̩̥̙̝ͅ ̲̠̬̥Be̡̙̫̦h̰̩i̛̫͙͔̭̤̗̲n̳͞d̸ ͎̻͘T̛͇̝̲̹̠̗ͅh̫̦̝ͅe̩̫͟ ͓͖̼W͕̳͎͚̙̥ą̙l̘͚̺͔͞ͅl̳͍̙̤̤̮̳.̢  ̟̺̜̙͉Z̤̲̙̙͎̥̝A͎̣͔̙͘L̥̻̗̳̻̳̳͢G͉̖̯͓̞̩̦O̹̹̺!̙͈͎̞̬ T̷̗͑̃ͦ̊͊̓͡͞h̛̥͚̖͇̝̬̹̟̃̋̽ͥ͆̂ͦ͗ȅ̱͗͘ ̱͈͋͛͡͝e͙͓̼̰̹̲̒ͤ́ͫ̓́͊͆͢n͂͗͊͌ͦ̐̊҉̧̟̺t̐ͧ̍̀ͤ̈́̔̒͏̻̜͙̝͕͍ͅï̛̠̩̦̿̉ͪ͌ͩ̚r̶̳̺͈̺͈͚͍ͯ̄ͤ͌̋͂͝e͖̟͓̝̩̺̬̭͙̐̉ ̵̴̴̗̤̺̥̰͚̐ͧͥͦ̿͛ͤr͙̣͔̺͆͞o̶̠̯͇͈ͧͭ̄ͩ̔̍͞ơ̖͙͔̥͋ͣ̈m͙͉̱̱̞̦̘̰͑ͭ͒̂̀͘ ̛̔ͭ̈ͬͮ҉̞͢ͅiͮ̒̔͘͏͉̯͎̦s̶̡͈͎̱̖͎̫̙̊ͫ̿̋ͣ͡ͅ ̙̣̬̟͈͋̓f͍͈̰̘̻̃͋͂̐i̸̺͙̪͔̞͙̣ͧͨ͝lͬ͂҉̸̻͔̙l͖̼ͮ͢͝e̷̟̙͔̠̯̓̓ͧ́ͬͪ̇̃d̶͕͇͎̦̐̓ ̮̯̥́͋̚̕w͑̍̔̔̀ͪͯ͘͢͏͉̰i̛̟̰̣ͭ̌̊͑̒ͫ̉͠t͇̖̬̠̗̲̄̓̈ͣͮ̂͂̊͗ḩ̲̖̊ͪ̓̄ ̝̺̟͕͖̈͋̎ͩͩ͆̈́̿͡Z̵̴̖͖͕̔ͬͮ͒̏̅̍̎a̤͖ͬ͑̎͐͑̔ͭ͞l̝̼̩͋̂g̨ͦ̓̓̏̈́̉ͯ͏͎͔̟̮̠̬͙ỏ͍̝̺͕͈ͭ̓̏̽̓ͭ̾́͢͞.͓̹͇̬̔̓̏ͦ̚͜͠" };
  34         const char *thai[] = { "ก็็็็็็็็็็กิิิิิิิิิิก้้้้้้้้้้ก็็็็็็็็็็ก็็็็็็็็็็กิิิิิิิิิิก้้้้้้้้้้" };
  35
  36         unsigned int zalgolen = strlen(*zalgo);
  37         unsigned int thailen = strlen(*thai);
  38
  39         char *ptr = page;
  40
  41         while (i < (page_size - zalgolen)) {
  42
  43                 j = rand() % 9;
  44
  45                 switch (j) {
  46
  47                 case 0:
  48                         strncpy(ptr, unicode1, 4);
  49                         ptr += 4;
  50                         i += 4;
  51                         break;
  52
  53                 case 1: unilen = rand() % 10;
  54                         for (l = 0; l < unilen; l++) {
  55                                 strncpy(ptr, unicode2, 6);
  56                                 ptr += 6;
  57                                 i += 6;
  58                                 if ((i + 6) > page_size)
  59                                         break;
  60                         }
  61                         break;
  62
  63                 case 2: strncpy(ptr, unicode3, 2);
  64                         i += 2;
  65                         ptr += 2;
  66                         break;
  67                 case 3: strncpy(ptr, unicode4, 18);
  68                         i += 18;
  69                         ptr += 18;
  70                         break;
  71
  72                 case 4: strncpy(ptr, unicode5, 4);
  73                         i += 4;
  74                         ptr += 4;
  75                         break;
  76
  77                 case 5: unilen = rand() % 10;
  78                         for (l = 0; l < unilen; l++) {
  79                                 strncpy(ptr, unicode6, 4);
  80                                 ptr += 4;
  81                                 i += 4;
  82                                 if ((i + 4) > page_size)
  83                                         break;
  84                         }
  85                         break;
  86
  87                 case 6: strncpy(ptr, unicode7, 4);
  88                         i += 4;
  89                         ptr += 4;
  90                         break;
  91
  92                 /* HE COMES. */
  93                 case 7: strncpy(ptr, *zalgo, zalgolen);
  94                         i += zalgolen;
  95                         ptr += zalgolen;
  96                         break;
  97
  98                 case 8: strncpy(ptr, *thai, thailen);
  99                         i += thailen;
 100                         ptr += thailen;
 101                         break;
 102
 103                 default:
 104                         break;
 105                 }
 106         }
 107
 108         page[rand() % page_size] = 0;
 109 }
 110
 111 #ifdef STANDALONE
 112 /*
 113  * gcc -I include -g -DSTANDALONE unicode.c -o unicode
 114  */
 115 #include <stdio.h>
 116 #include <stdlib.h>
 117 #include <string.h>
 118
 119 unsigned int page_size = 4096;
 120
 121 void main(int argc, char* argv[])
 122 {
 123         unsigned char *page;
 124         unsigned int x = 0, y, n = 0;
 125         struct timeval t;
 126
 127         gettimeofday(&t, 0);
 128         srand((t.tv_sec * getpid()) ^ t.tv_usec);
 129
 130         page = malloc(4096);
 131         memset(page, 0, 4096);
 132
 133         gen_unicode_page(page);
 134
 135         for (y = 0; y < 4096; y+=32) {
 136                 for (x = 0; x < 32; x++) {
 137                         printf("%c", page[n++]);
 138                 }
 139         }
 140 }
 141 #endif