/*
 *  List up recently trackbacks
 *
 *    Copyright (c) 2004 by H.Tsujimura (tsupo@na.rim.or.jp)
 *    All Rights Reserved.
 *
 *      written by H.Tsujimura    28 Feb 2004
 *      Last update:              02 Mar 2004
 *
 *        See Also:
 *          http://www.na.rim.or.jp/~tsupo/program/
 *          http://watcher.moe-nifty.com/
 */

/*
 *  Usage:
 *     listTB [-p] [-v] [-nNN] Target Blog's URL
 *         -p: get html pages via proxy
 *         -v: verbose mode
 *         -n: number of trackbacks for listing
 *              (default: 15 trackbacks)
 *     ex)
 *         listTB -n30 http://watcher.moe-nifty.com/memo/
 *
 */

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#ifdef WIN32
#include <time.h>
#else
#include <sys/time.h>
#endif

struct trackbackInfo {
    char    trackbackEntryURL[256];    /* トラックバック元URL      */
    char    trackbackEntryTitle[256];  /* トラックバック元記事題名 */
    char    trackbackedURL[256];       /* 被トラックバックURL      */
    char    trackbackedEntryTitle[256];/* 被トラックバック記事題名 */
#if 0
    int     yy; /* トラックバック受信 年 */
    int     mm; /* トラックバック受信 月 */
    int     dd; /* トラックバック受信 日 */
    int     HH; /* トラックバック受信 時 */
    int     MM; /* トラックバック受信 分 */
    int     SS; /* トラックバック受信 秒 */
#endif
}       *tbList = NULL;
long    numOfTbList  = 0;
long    numOfTables  = 0;
long    numOfDisplay = 15;

char    *
httpGet( const char *url, int useProxy, int verbose )
{
    static char filename[128];
    char        cmd[BUFSIZ];
    char        *p, *q;

    p = strrchr( url, '.' );
    q = strrchr( url, '/' );
    if ( p && q && (q < p) ) {
        strncpy( filename, q + 1, p - url );
        strcpy( &(filename[p - url]), "_tbChk.html" );
    }
    else
        strcpy( filename, "tbChk.html" );

    sprintf( cmd, "httpRead%s%s > %s",
             useProxy && verbose ? " -pv " :
             useProxy            ? " -p "  :
             verbose             ? " -v "  :
                                   " ",
             url,
             filename );
    if ( verbose )
        fprintf( stderr, "%% %s\n", cmd );
    system( cmd );
        /* 'httpRead' is a program for get a html page */
        /* via HTTP GET method.                        */ 

    return ( filename );
}


void
listupTrackbackUrls( const char *filename, int verbose )
{
    FILE                    *fp;
    char                    buf[BUFSIZ], *p, *q;
    int                     flag = 0;
    struct trackbackInfo    tbInfo, *tp, *tq;

    memset( &tbInfo, 0x00, sizeof ( struct trackbackInfo ) );
    if ( !tbList ) {
        tbList = (struct trackbackInfo *)
                    malloc( 1024 * sizeof ( struct trackbackInfo ) );
        if ( !tbList )
            return;
        numOfTables++;
    }

    tp = &tbInfo;
    if ( ( fp = fopen( filename, "r" ) ) != NULL ) {
        while ( ( p = fgets( buf, BUFSIZ - 1, fp ) ) != NULL ) {
            while ( (*p == ' ') || (*p == '\t') )
                p++;
            if ( flag == 0 ) {
                if ( ( !strncmp( p, "<title>", 7 ) ) ) {
                    p += 7;
                    q = strstr( p, "</title>" );
                    strncpy( tp->trackbackedEntryTitle, p, q - p );
                    tp->trackbackedEntryTitle[q - p] = '\0';
                }
                else if ( ( !strncmp( p, "<link>", 6 ) ) ) {
                    p += 6;
                    q = strstr( p, "</link>" );
                    strncpy( tp->trackbackedURL, p, q - p );
                    tp->trackbackedURL[q - p] = '\0';
                }
                else if ( ( q = strstr( p, "<item>" ) ) != NULL )
                    flag++;
                continue;
            }

            if ( ( !strncmp( p, "<title>", 7 ) ) ) {
                p += 7;
                q = strstr( p, "</title>" );
                strncpy( tp->trackbackEntryTitle, p, q - p );
                tp->trackbackEntryTitle[q - p] = '\0';
                flag++;
            }
            else if ( ( !strncmp( p, "<link>", 6 ) ) ) {
                p += 6;
                q = strstr( p, "</link>" );
                strncpy( tp->trackbackEntryURL, p, q - p );
                tp->trackbackEntryURL[q - p] = '\0';
                flag++;
            }

            if ( flag >= 3 ) {
                tq = &(tbList[numOfTbList]);
                memcpy( tq, tp, sizeof( struct trackbackInfo ) );

                if ( verbose )
                    fprintf( stderr, "%4d) %s <- %s\n",
                             numOfTbList,
                             tq->trackbackedURL, 
                             tq->trackbackEntryURL );

                numOfTbList++;
                flag = 1;
                tp->trackbackEntryTitle[0] = '\0';
                tp->trackbackEntryURL[0]   = '\0';
                
                if ( numOfTbList >= numOfTables * 1024 ) {
                    struct trackbackInfo *p;

                    p = (struct trackbackInfo *)
                        realloc( tbList,
                                 numOfTables * 1024 *
                                     sizeof ( struct trackbackInfo ) );
                    if ( !p )
                        return;
                    tbList = p;
                    numOfTables++;
                }
            }
        }
        fclose( fp );
    }
}


void
listTB( const char *url,    /* top page's url   */
        int        useProxy,/* use proxy or not */
        int        verbose  /* use verbose mode */
      )
{
    char    *filename;
    char    buf[BUFSIZ];
    char    *p;
    int     flag = 0;
    int     done = 0;
    char    topEntryFileName[256];
    char    prevPageFileName[256];
    char    trackBackUrl[256];
    FILE    *fp;

    topEntryFileName[0] = '\0';

    /* トップページを取得 */
    filename = httpGet( url, useProxy, verbose );

    /* トップページを解析         */
    /* 先頭の記事ファイル名を取得 */
    if ( ( fp = fopen( filename, "r" ) ) != NULL ) {
        while ( ( p = fgets( buf, BUFSIZ - 1, fp ) ) != NULL ) {
            while ( (*p == ' ') || (*p == '\t') )
                p++;
            if ( (flag == 0) && !strncmp( p, "<rdf:RDF", 8 ) ) {
                flag++;
                continue;
            }
            if ( (flag == 1) && !strncmp( p, "rdf:about=\"", 11 ) ) {
                strcpy( topEntryFileName, p + 11 );
                if ( (p = strrchr( topEntryFileName, '"' )) != NULL )
                    *p = '\0';
                break;
            }
        }
        fclose( fp );
        unlink( filename );
    }

    /* 先頭の記事を取得 */
    if ( topEntryFileName[0] ) {
        filename = httpGet( topEntryFileName, useProxy, verbose );

        do {
            /* 1つ前の記事とトラックバックの URL を各々取得 */
            flag = 0;
            prevPageFileName[0] = '\0';
            trackBackUrl[0]     = '\0';
            if ( ( fp = fopen( filename, "r" ) ) != NULL ) {
                while ( ( p = fgets( buf, BUFSIZ - 1, fp ) ) != NULL ) {
                    while ( (*p == ' ') || (*p == '\t') )
                        p++;
                    if ( (flag == 0) && !strncmp( p, "<title>", 7 ) ) {
                        flag++;
                        continue;
                    }
                    if ( (flag == 1) &&
                         !strncmp( p, "<link rel=\"prev\" href=\"", 23 ) ) {
                        strcpy( prevPageFileName, p + 23 );
                        if ( (p = strchr( prevPageFileName, '"' )) != NULL )
                            *p = '\0';
                        flag++;
                        continue;
                    }
                    if ( (flag == 2) &&
                         !strncmp( p, "trackback:ping=\"", 16 ) ) {
                        strcpy( trackBackUrl, p + 16 );
                        if ( (p = strchr( trackBackUrl, '"' )) != NULL )
                            *p = '\0';
                        flag++;
                        continue;
                    }
                }
                fclose( fp );
                unlink( filename );
            }

            if ( trackBackUrl[0] ) {
                /* トラックバック元URL一覧を取得 */
                char    tbUrl[256];
                strcat( trackBackUrl, "?__mode=rss" );
                filename = httpGet( trackBackUrl, useProxy, verbose );
                listupTrackbackUrls( filename, verbose );
                unlink( filename );

                if ( (numOfDisplay > 0) && (numOfTbList > numOfDisplay) )
                    break;
            }
            if ( prevPageFileName[0] ) {
                /* 1つ前の記事の処理へ進む */
                filename = httpGet( prevPageFileName, useProxy, verbose );
            }
            else
                done = 1;
        } while ( !done );
    }

    if ( numOfTbList> 0 ) {
        int i, flag = 0;
        printf( "<html><head><meta http-equiv=\"Content-Type\" "
                "content=\"text/html; charset=UTF-8\" />\n" );
        printf( "<title>List of Trackbacked URL</title>\n" );
        printf( "</head>\n\n<body>\n" );
        printf( "<h2>&#12488;&#12521;&#12483;&#12463;&#12496;"
                "&#12483;&#12463;&#19968;&#35239;</h2>\n<ul>\n" );
        for ( i = 0; i < numOfTbList; i++ ) {
            if ( (i == 0) ||
                 strcmp( tbList[i].trackbackedURL,
                         tbList[i-1].trackbackedURL ) ) {
                if ( flag == 1 )
                    printf( "</ul></li>\n" );
                printf( "<li><a href=\"%s\">%s</a>\n<ul><li>"
                        "<a href=\"%s\">%s</a></li>\n",
                        tbList[i].trackbackedURL,
                        tbList[i].trackbackedEntryTitle,
                        tbList[i].trackbackEntryURL,
                        tbList[i].trackbackEntryTitle );
                flag = 1;
            }
            else {
                printf( "<li><a href=\"%s\">%s</a></li>\n",
                        tbList[i].trackbackEntryURL,
                        tbList[i].trackbackEntryTitle );
                flag = 1;
            }
        }
        if ( flag == 1 )
            printf( "</ul></li>\n" );
        printf( "</ul>\n</body>\n</html>\n\n" );
    }
}

int
main( int argc, char *argv[] )
{
    int     i;
    int     verbose  = 0;
    int     useProxy = 0;
    char    *p;

    if ( argc <= 1 ) {
        fprintf( stderr, "listTB: too few arguments.\n" );
        return ( 1 );
    }

    for ( i = 1; i < argc; i++ ) {
        if ( argv[i][0] == '-' ) {
            for ( p = &(argv[i][1]); *p; p++ ) {
                switch( *p ) {
                case 'v':
                    verbose = !verbose;
                    break;
                case 'p':
                    useProxy = !useProxy;
                    break;
                case 'n':
                    if ( (*(p + 1) >= '0') && (*(p + 1) <= '9') )
                        numOfDisplay = atol( p + 1 );
                    break;
                }
            }
            continue;
        }

        listTB( argv[i], useProxy, verbose );
    }

    return ( 0 );
}