API tools faq
paste
Guest User

Untitled

a guest
Jul 16th, 2018
72
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 8.92 KB | None | 0 0
raw download clone embed print report
  1. /* Xingbo Wang CSC192 997583155 Assignment 3 */
  2.  
  3. #define _GNU_SOURCE
  4. #include <stdlib.h>
  5. #include <stdio.h>
  6. #include <string.h>
  7.  
  8. #define TAGFILENAME "tags.txt"
  9. #define TAGSTOCHECK 23
  10. #define MAXTAGNAMESIZE 25
  11.  
  12. typedef struct __tagtype{
  13. char name[MAXTAGNAMESIZE];
  14. unsigned short closing; //closing tag required? 1 = yes, 0 = no, 2 = maybe
  15. unsigned short depend; //must be used inside which tag?
  16. unsigned short opened; //count of num of times tag is opened
  17. unsigned short closed; //count of num of times tag is closed
  18. } tagtype;
  19.  
  20. unsigned int error = 0;
  21. unsigned int linen = 0, tagn = 0;
  22. unsigned int currenttag = 0, lasttag = 0;
  23. tagtype * tags = NULL;
  24. char * tempstr = NULL; //storage of temporary string to avoid memory leak
  25. char * string = NULL;
  26. char * tagstr = NULL;
  27. FILE * filep = NULL;
  28. FILE * tagfile = NULL;
  29.  
  30. //Waste of memory, only needed for one of the assignment's ridiculous requirements
  31. char realname1[MAXTAGNAMESIZE], realname2[MAXTAGNAMESIZE];
  32.  
  33.  
  34. /*----------------------------------------------------------------------------*/
  35.  
  36. int sumopentags(void)
  37. /* returns the sum of the number of currently unclosed tags */
  38. {
  39. int sum = 0, i = 0;
  40. for(i = 1; i < TAGSTOCHECK; i++) sum += (tags[i].opened - tags[i].closed);
  41. return sum;
  42. }
  43.  
  44. /*----------------------------------------------------------------------------*/
  45.  
  46. unsigned int whichtag(char * tag)
  47. /* returns tagtype index of the matching tag name, 0 if not found in tagtype */
  48. {
  49. unsigned int i = 0;
  50. for(i = 1; i < tagn; i++){
  51. if(strcmp(tag, tags[i].name) == 0){
  52. return i;
  53. }
  54. }
  55. return 0;
  56. }
  57.  
  58. /*----------------------------------------------------------------------------*/
  59.  
  60. unsigned int checkerrors()
  61. /* check for the 9 errors listed below
  62. 1 "<html> not first tag of document"
  63. 2 "</html> not last tag of document"
  64. 3 "<head> not first tag of html block"
  65. 4 "</body> not last tag of html block"
  66. 5 "html block does not contain exactly 1 head block"
  67. 6 "html block does not contain exactly 1 body block"
  68. 7 "tag used outside of valid block on line Z"
  69. 8 "closing </img> found on line Z"
  70. 9 "tags not properly closed; expecting </X>, got </Y> on line %u"
  71. */
  72. {
  73. int sum = sumopentags();
  74.  
  75. if(tags[whichtag("html")].opened < 1 && (sum == 1 || (sum == 0 && tags[currenttag].opened > 0 && currenttag != 0))) return 1;
  76. if(tags[whichtag("html")].closed > 0 && (sum > 0 || (sum == 0 && tags[currenttag].closing != 1))) return 2;
  77. if(tags[whichtag("head")].opened < 1 && (sum > 1 || (sum > 0 && tags[currenttag].closing != 1)) && currenttag != 0) return 3;
  78. if(tags[whichtag("body")].closed > 0 && tags[whichtag("html")].closed < 1 &&
  79. (sum > 1 || (sum == 1 && tags[currenttag].closing != 1))) return 4;
  80. if(tags[whichtag("head")].opened > 1 || (tags[whichtag("html")].closed >= 1 && tags[whichtag("head")].opened < 1)) return 5;
  81. if(tags[whichtag("head")].opened > 1 || (tags[whichtag("html")].closed >= 1 && tags[whichtag("head")].opened < 1)) return 6;
  82.  
  83. //special case for <li>
  84. if(currenttag == whichtag("li") && (tags[whichtag("ol")].opened - tags[whichtag("ol")].closed) == 0
  85. && (tags[whichtag("ul")].opened - tags[whichtag("ul")].closed) == 0) return 7;
  86. if((tags[currenttag].opened - tags[currenttag].closed) >= 1 && currenttag != whichtag("li")){
  87. if(tags[tags[currenttag].depend].opened < 1 ||
  88. (tags[tags[currenttag].depend].opened - tags[tags[currenttag].depend].closed) == 0)
  89. return 7;
  90. }else if (tags[currenttag].closing != 1 && currenttag != whichtag("li")){
  91. if(tags[tags[currenttag].depend].opened < 1 ||
  92. ((tags[tags[currenttag].depend].opened - tags[tags[currenttag].depend].closed) == 0
  93. && tags[tags[currenttag].depend].closing == 1))
  94. return 7;
  95. }
  96.  
  97. if(tags[whichtag("img")].closed > tags[whichtag("img")].opened) return 8;
  98. if((tags[lasttag].opened - tags[lasttag].closed) >= 1 &&
  99. (tags[currenttag].opened - tags[currenttag].closed) == 0 &&
  100. tags[currenttag].closing == 1)
  101. return 9;
  102.  
  103. //no errors
  104. return 0;
  105. }
  106.  
  107. /*----------------------------------------------------------------------------*/
  108.  
  109. void printerror(int errnum)
  110. /* prints the corresponding error string to stderr */
  111. {
  112. switch(errnum){
  113. case 1 : fprintf(stderr, "<html> not first tag of document\n"); break;
  114. case 2 : fprintf(stderr, "</html> not last tag of document\n"); break;
  115. case 3 : fprintf(stderr, "<head> not first tag of html block\n"); break;
  116. case 4 : fprintf(stderr, "</body> not last tag of html block\n"); break;
  117. case 5 : fprintf(stderr, "html block does not contain exactly 1 head block\n"); break;
  118. case 6 : fprintf(stderr, "html block does not contain exactly 1 body block\n"); break;
  119. case 7 : fprintf(stderr, "tag used outside of valid block on line %u\n", linen); break;
  120. case 8 : fprintf(stderr, "closing </img> found on line %u\n", linen); break;
  121. case 9 : fprintf(stderr, "tags not properly closed; expecting </%s>, got </%s> on line %u\n",
  122. realname2, realname1, linen); break;
  123. }
  124. }
  125.  
  126. /*----------------------------------------------------------------------------*/
  127.  
  128. void lowercase(char * string, unsigned int len)
  129. /* makes all the letters in a string lowercase */
  130. {
  131. unsigned int i = 0;
  132. for(i = 0; i < len; i++)
  133. if(string[i] <= 90 && string[i] >= 65)
  134. string[i] += 32;
  135. }
  136.  
  137. /*----------------------------------------------------------------------------*/
  138.  
  139. char * gettag(char * tagstr, unsigned int len)
  140. /* returns the tag name, e.g. "body" from <BODY background....> , retains original
  141. case of the tag name in realname1 and realname2*/
  142. {
  143. /* extract the first word*/
  144. tempstr = (char*)calloc(len, sizeof(char));
  145. memset(tempstr, 0, len);
  146. strncpy(tempstr, tagstr, (int)(strchrnul(tagstr, ' ') - tagstr));
  147. strcpy(realname2, realname1);
  148. strncpy(realname1, tempstr, MAXTAGNAMESIZE);
  149. lowercase(tempstr, len);
  150. return tempstr;
  151. }
  152.  
  153. /*----------------------------------------------------------------------------*/
  154.  
  155. int parseline(char * string, unsigned int len)
  156. /* parses a single line */
  157. {
  158. unsigned int n = 0, start = 0;
  159. tagstr = (char*)calloc(len, sizeof(char));
  160.  
  161. for(n = 0; n < len; n++){
  162. memset(tagstr, 0, len);
  163. if(string[n] == '<'){
  164. start = n;
  165. }else if(string[n] == '>'){
  166. /* parse the tag */
  167. /* record which tag was opened */
  168. strncpy(tagstr, &(string[start+1]), (n-start-1)*sizeof(char));
  169. if(tagstr[0] == '/'){
  170. currenttag = whichtag(gettag(&tagstr[1], (n-start)));
  171. if(currenttag > 0 && tags[currenttag].closing == 1){
  172. tags[currenttag].closed += 1;
  173. }
  174. }else{
  175. currenttag = whichtag(gettag(tagstr, (n-start)));
  176. if(currenttag > 0){
  177. tags[currenttag].opened += 1;
  178. /* close it too if closing not required */
  179. if(tags[currenttag].closing != 1) tags[currenttag].closed += 1;
  180. }
  181. }
  182. /* check for errors, print and return if found */
  183. error = checkerrors();
  184. if(error != 0) return 1;
  185. }
  186. if(currenttag != 0) lasttag = currenttag;
  187. }
  188. return 0;
  189. }
  190.  
  191. /*----------------------------------------------------------------------------*/
  192.  
  193. void free_all()
  194. /* frees all of the pointers and files */
  195. {
  196. if (string != NULL) free(string);
  197. if (tempstr != NULL) free(tempstr);
  198. if (tagstr != NULL) free(tagstr);
  199. if (filep != NULL) fclose(filep);
  200. if (tagfile != NULL) fclose(tagfile);
  201. if (tags != NULL) free(tags);
  202. }
  203.  
  204. /*----------------------------------------------------------------------------*/
  205.  
  206. int main(int argc, char * argv[])
  207. {
  208. /* open files, check for file errors */
  209. if(argc != 2){
  210. fprintf(stderr,"Wrong number of arguments\n");
  211. return 1;
  212. }
  213. filep = fopen(argv[1], "r");
  214. if(filep == NULL){
  215. fprintf(stderr, "ERROR %s not found\n", argv[1]);
  216. free_all();
  217. return EXIT_FAILURE;
  218. }
  219. tagfile = fopen(TAGFILENAME, "r");
  220. if(tagfile == NULL){
  221. fprintf(stderr, "ERROR %s not found\n", TAGFILENAME);
  222. free_all();
  223. return EXIT_FAILURE;
  224. }
  225.  
  226. unsigned int n = 0;
  227.  
  228. /* set up tagtype, match up index numbers with tags.txt line numbers */
  229. tags = (tagtype*) calloc(TAGSTOCHECK, sizeof(tagtype));
  230. tags[0].opened = 1; //<html> depends on this to be open
  231. tagn++;
  232.  
  233. /* read in the tag file and store their properties into tagtype*/
  234. do{
  235. string = (char*) realloc(string, ++n*sizeof(char));
  236. string[n-1] = fgetc(tagfile);
  237. if(string[n-1] == '\n'){
  238. tagn++;
  239. sscanf(string, "%s %hu %hu", tags[tagn-1].name, &tags[tagn-1].closing,
  240. &tags[tagn-1].depend);
  241. string = NULL;
  242. n = 0;
  243. }
  244. }while(!feof(tagfile));
  245.  
  246. /* read html and parse */
  247. do{
  248. string = (char*) realloc(string, ++n*sizeof(char));
  249. string[n-1] = fgetc(filep);
  250. if(string[n-1] == '\n'){
  251. linen++;
  252. if(parseline(string, n) == 1) break;
  253. string = NULL;
  254. n = 0;
  255. }
  256. }while(!feof(filep));
  257.  
  258. /* evaluate */
  259. if (currenttag != 1 && error == 0) error = 2; //if eof and </html> not found
  260. if (error != 0){
  261. printf("Verified. Page is invalid.\n");
  262. fprintf(stderr, "REASON ");
  263. printerror(error);
  264. }
  265. else printf("Verified. Page is valid.\n");
  266.  
  267. /* finish */
  268. free_all();
  269.  
  270. return EXIT_SUCCESS;
  271. }
Add Comment
Please, Sign In to add comment
Public Pastes
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy.  OK, I Understand
Not a member of Pastebin yet?
Sign Up, it unlocks many cool features!