CMakeLua/Utilities/cmcurl-7.19.0/docs/examples/htmltidy.c

   1 /*****************************************************************************
   2  *                                  _   _ ____  _
   3  *  Project                     ___| | | |  _ \| |
   4  *                             / __| | | | |_) | |
   5  *                            | (__| |_| |  _ <| |___
   6  *                             \___|\___/|_| \_\_____|
   7  *
   8  * $Id: htmltidy.c,v 1.1.1.1 2008-09-23 16:32:05 hoffman Exp $
   9  *
  10  * Download a document and use libtidy to parse the HTML.
  11  * Written by Jeff Pohlmeyer
  12  *
  13  * LibTidy => http://tidy.sourceforge.net
  14  *
  15  * gcc -Wall -I/usr/local/include tidycurl.c -lcurl -ltidy -o tidycurl
  16  *
  17  */
  18
  19 #include <stdio.h>
  20 #include <tidy/tidy.h>
  21 #include <tidy/buffio.h>
  22 #include <curl/curl.h>
  23
  24 /* curl write callback, to fill tidy's input buffer...  */
  25 uint write_cb(char *in, uint size, uint nmemb, TidyBuffer *out)
  26 {
  27   uint r;
  28   r = size * nmemb;
  29   tidyBufAppend( out, in, r );
  30   return(r);
  31 }
  32
  33 /* Traverse the document tree */
  34 void dumpNode(TidyDoc doc, TidyNode tnod, int indent )
  35 {
  36   TidyNode child;
  37   for ( child = tidyGetChild(tnod); child; child = tidyGetNext(child) )
  38   {
  39     ctmbstr name = tidyNodeGetName( child );
  40     if ( name )
  41     {
  42       /* if it has a name, then it's an HTML tag ... */
  43       TidyAttr attr;
  44       printf( "%*.*s%s ", indent, indent, "<", name);
  45       /* walk the attribute list */
  46       for ( attr=tidyAttrFirst(child); attr; attr=tidyAttrNext(attr) ) {
  47         printf(tidyAttrName(attr));
  48         tidyAttrValue(attr)?printf("=\"%s\" ",
  49                                    tidyAttrValue(attr)):printf(" ");
  50       }
  51       printf( ">\n");
  52     }
  53     else {
  54       /* if it doesn't have a name, then it's probably text, cdata, etc... */
  55       TidyBuffer buf;
  56       tidyBufInit(&buf);
  57       tidyNodeGetText(doc, child, &buf);
  58       printf("%*.*s\n", indent, indent, buf.bp?(char *)buf.bp:"");
  59       tidyBufFree(&buf);
  60     }
  61     dumpNode( doc, child, indent + 4 ); /* recursive */
  62   }
  63 }
  64
  65
  66 int main(int argc, char **argv )
  67 {
  68   CURL *curl;
  69   char curl_errbuf[CURL_ERROR_SIZE];
  70   TidyDoc tdoc;
  71   TidyBuffer docbuf = {0};
  72   TidyBuffer tidy_errbuf = {0};
  73   int err;
  74   if ( argc == 2) {
  75     curl = curl_easy_init();
  76     curl_easy_setopt(curl, CURLOPT_URL, argv[1]);
  77     curl_easy_setopt(curl, CURLOPT_ERRORBUFFER, curl_errbuf);
  78     curl_easy_setopt(curl, CURLOPT_NOPROGRESS, 0L);
  79     curl_easy_setopt(curl, CURLOPT_VERBOSE, 1L);
  80     curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, write_cb);
  81
  82     tdoc = tidyCreate();
  83     tidyOptSetBool(tdoc, TidyForceOutput, yes); /* try harder */
  84     tidyOptSetInt(tdoc, TidyWrapLen, 4096);
  85     tidySetErrorBuffer( tdoc, &tidy_errbuf );
  86     tidyBufInit(&docbuf);
  87
  88     curl_easy_setopt(curl, CURLOPT_WRITEDATA, &docbuf);
  89     err=curl_easy_perform(curl);
  90     if ( !err ) {
  91       err = tidyParseBuffer(tdoc, &docbuf); /* parse the input */
  92       if ( err >= 0 ) {
  93         err = tidyCleanAndRepair(tdoc); /* fix any problems */
  94         if ( err >= 0 ) {
  95           err = tidyRunDiagnostics(tdoc); /* load tidy error buffer */
  96           if ( err >= 0 ) {
  97             dumpNode( tdoc, tidyGetRoot(tdoc), 0 ); /* walk the tree */
  98             fprintf(stderr, "%s\n", tidy_errbuf.bp); /* show errors */
  99           }
 100         }
 101       }
 102     }
 103     else
 104       fprintf(stderr, "%s\n", curl_errbuf);
 105
 106     /* clean-up */
 107     curl_easy_cleanup(curl);
 108     tidyBufFree(&docbuf);
 109     tidyBufFree(&tidy_errbuf);
 110     tidyRelease(tdoc);
 111     return(err);
 112
 113   }
 114   else
 115     printf( "usage: %s <url>\n", argv[0] );
 116
 117   return(0);
 118 }