#include "headers.h"


/* really hate using the preprocessor, but it makes sense in this context   */
/* SUBSTR_COUNT is number of parenthetical substrings in REGEX_URI plus one */
/* this regex is from the RFC describing URI syntax -- can't recall the     */
/* exact one right now. anyway, it's a little too general for my tastes,    */
/* but the one I came up with was trash (unsurprisingly) so here we are.    */
/* need to modify this in the future to be less liberal...                  */
#define REGEX_URI "^([^:/?#]+)://(([^/?#]*)+([^?#]*))(\\?([^#]*))?(#(.*))?"


#define SUBSTR_COUNT 8
#define PROTO 1
#define FQDN  3
#define PATH  4


int uri_parse(const char *uristr, uri *res)
  {
  char validp;
  regex_t regexp;
  regmatch_t match[SUBSTR_COUNT];

  validp = 0;
  

  regcomp(&regexp, REGEX_URI, REG_EXTENDED);
  if( (validp = regexec(&regexp, uristr, SUBSTR_COUNT, match, 0)))
    {
    return(1);
    }


  /* not very elegant but it does the job. i shouldn't be thinking about */
  /* elegance at this stage in my programming life, anyways... comes     */
  /* with experience.                                                    */
  res->proto = substr_extract(uristr, match[PROTO].rm_so, match[PROTO].rm_eo);
  res->fqdn  = substr_extract(uristr, match[FQDN].rm_so, match[FQDN].rm_eo);

  /* if the difference below is less than 1, our path doesn't exist. */
  /* Compensate by setting it to '/' which will always return a root */
  /* document from an HTTP server -- and, presumably, others. We'll  */
  /* see, I suppose.                                                 */
  if((match[PATH].rm_eo - match[PATH].rm_so) < 1)
      {
      res->path = "/";
      }
  else
    {
    res->path  = substr_extract(uristr, match[PATH].rm_so, match[PATH].rm_eo);
    }


  return(0);
  }