%{								/* -*- C -*- */
# include <stdio.h>
# include <curl/curl.h>

  typedef struct StringList
  {
    char		*string;
    struct StringList	*next;
  } StringList;

  StringList	*newString(char *string, StringList *list);
  void		 pushString(StringList **list, char *string);
  char		*popString(StringList **list);

  StringList *urls= 0;

  FILE *yyin= 0;

# define YY_INPUT(buf, result, max_size)		\
  {							\
    int yyc= fgetc(yyin);				\
    result= (EOF == yyc) ? 0 : (*(buf)= yyc, 1);	\
  }
%}

document = (href | .)* !.

href = '<' - [aA] (!HREF .)* HREF - '=' - '"' < (!'"' .)+ > '"' (!'>' .)* '>'		{ pushString(&urls, yytext); }

HREF	= [hH][rR][eF][fF]

- = (' ' | '\t' | '\n' | '\r')*

%%

int verbose= 0;

StringList *errorFiles= 0;
StringList *errorURLs= 0;
StringList *errorMessages= 0;

StringList *newString(char *string, StringList *list)
{
  StringList *s= calloc(1, sizeof(StringList));
  s->string= strdup(string);
  s->next= list;
  return s;
}

void pushString(StringList **list, char *string)
{
  *list= newString(string, *list);
}

char *popString(StringList **list)
{
  StringList *s= *list;
  char *string;
  if (!s)
    {
      fprintf(stderr, "string list underflow\n");
      exit(1);
    }
  *list= s->next;
  string= s->string;
  free(s);
  return string;
}

void error(char *file, char *url, char *message)
{
  pushString(&errorFiles,    file);
  pushString(&errorURLs,     url);
  pushString(&errorMessages, message);
  fprintf(stderr, "%s: %s: %s\n", file, message, url);
}

void checklink(char *file, char *url)
{
  if ('#' != url[0])
    {
      CURL *curl= curl_easy_init();
      CURLcode code;
      long response;
      curl_easy_setopt(curl, CURLOPT_URL,	url);
      curl_easy_setopt(curl, CURLOPT_TIMEOUT,	30);
      curl_easy_setopt(curl, CURLOPT_NOBODY,	1);
      if ((code= curl_easy_perform(curl)))
	error(file, url, (char *)curl_easy_strerror(code));
      else
	{
	  if ((code= curl_easy_getinfo(curl, CURLINFO_RESPONSE_CODE, &response)))
	    error(file, url, (char *)curl_easy_strerror(code));
	  else
	    {
	      if (verbose)
		printf("%ld %s\n", response, url);
	      if (404 == response)
		error(file, url, "ERROR 404");
	    }
	}
      curl_easy_cleanup(curl);
    }
}

void usage(char *program)
{
  fprintf(stderr, "usage: %s [option] filename...\n", program);
  fprintf(stderr, "  -v     show response code for every URL\n");
}

int main(int argc, char **argv)
{
  int i;

  for (i= 1;  i < argc;  ++i)
    {
      if ('-' == !strncmp(argv[i], "-h", 2) || !strncmp(argv[i], "--h", 3))
	usage(argv[0]);
      else if (!strcmp(argv[i], "-v"))
	verbose= 1;
      else if (!(yyin= fopen(argv[i], "r")))
	perror(argv[i]);
      else
	{
	  fprintf(stderr, "checking %s\n", argv[i]);
	  yyparse();
	  if (!feof(yyin))
	    {
	      error(argv[i], "?", "parse error");
	      fprintf(stderr, "parse error at line %d\n");
	    }
	  else
	    while (urls)
	      checklink(argv[i], popString(&urls));
	  fclose(yyin);
	}
    }

  if (verbose && errorFiles)
    {
      fprintf(stderr, "\nErrors detected:\n");
      while (errorFiles)
	fprintf(stderr, "%s: %s: %s\n", popString(&errorFiles), popString(&errorMessages), popString(&errorURLs));
    }

  return 0;
}