path: root/innbbsd/str_decode.c

/* 使用方法大致如下: innbbsd 中
   在 SUBJECT 從 news 讀進來後, 呼叫 str_decode_M3(SUBJECT) 就行了 */

/* bsd 底下使用要編譯時要加  -I/usr/local/include -L/usr/local/lib -liconv
   並安裝 libiconv,
   若真的沒有iconv就把底下的 #define USE_ICONV 1 刪了 */

/*-------------------------------------------------------*/
/* lib/str_decode.c ( NTHU CS MapleBBS Ver 3.00 )    */
/*-------------------------------------------------------*/
/* target : included C for QP/BASE64 decoding        */
/* create : 95/03/29                     */
/* update : 97/03/29                     */
/*-------------------------------------------------------*/


/* ----------------------------------------------------- */
/* QP code : "0123456789ABCDEF"              */
/* ----------------------------------------------------- */

#include <stdio.h>
#include <errno.h>
#include <string.h>

#define USE_ICONV 1
/* bsd 底下使用要編譯時要加  -I/usr/local/include -L/usr/local/lib -liconv
   若真的沒有iconv就把上面這行 #define 刪了 */

#ifdef USE_ICONV
#include <iconv.h>
#endif

static int qp_code(int x)
{
    if (x >= '0' && x <= '9')
    return x - '0';
    if (x >= 'a' && x <= 'f')
    return x - 'a' + 10;
    if (x >= 'A' && x <= 'F')
    return x - 'A' + 10;
    return -1;
}


/* ------------------------------------------------------------------ */
/* BASE64 :                               */
/* "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/" */
/* ------------------------------------------------------------------ */


static int base64_code(int x)
{
    if (x >= 'A' && x <= 'Z')
    return x - 'A';
    if (x >= 'a' && x <= 'z')
    return x - 'a' + 26;
    if (x >= '0' && x <= '9')
    return x - '0' + 52;
    if (x == '+')
    return 62;
    if (x == '/')
    return 63;
    return -1;
}


/* ----------------------------------------------------- */
/* judge & decode QP / BASE64                */
/* ----------------------------------------------------- */

static inline int isreturn(unsigned char c)
{
    return c == '\r' || c == '\n';
}

static inline int isspace(unsigned char c)
{
    return c == ' ' || c == '\t' || isreturn(c);
}

/* static inline */
int mmdecode(unsigned char *src, unsigned char encode, unsigned char *dst)
{
    /* Thor.980901: src和dst可相同, 但src 一定有?或\0結束 */
    /* Thor.980901: 注意, decode出的結果不會自己加上 \0 */
    unsigned char *t = dst;
    int pattern = 0, bits = 0;
    encode |= 0x20;        /* Thor: to lower */
    switch(encode)
    {
    case 'q':            /* Thor: quoted-printable */
        while(*src && *src != '?') /* Thor: delimiter */
        {                  /* Thor.980901: 0 算是 delimiter */
        if(*src == '=')
        {
            int x = *++src, y = x ? *++src : 0;
            if(isreturn(x)) continue;
            if( (x = qp_code(x)) < 0 || ( y = qp_code(y)) < 0) return -1;
            *t++ = (x << 4) + y , src++;
        }
        else if(*src == '_')
            *t++ = ' ', src++;
#if 0
        else if(!*src)    /* Thor: no delimiter is not successful */
            return -1;
#endif
        else             /* Thor: *src != '=' '_' */
            *t++ = *src++;
        }
        return t - dst;
    case 'b':            /* Thor: base 64 */
        while(*src && *src != '?') /* Thor: delimiter */ /* Thor.980901: 0也算 */
        {                  /* Thor: pattern & bits are cleared outside */
        int x;
#if 0
        if (!*src) return -1; /* Thor: no delimiter is not successful */
#endif
        x = base64_code(*src++);
        if(x < 0) continue; /* Thor: ignore everything not in the base64,=,.. */
        pattern = (pattern << 6) | x;
        bits += 6;       /* Thor: 1 code gains 6 bits */
        if(bits >= 8)    /* Thor: enough to form a byte */
        {
            bits -= 8;
            *t++ = (pattern >> bits) & 0xff;
        }
        }
        return t - dst;
    }
    return -1;
}

#ifdef USE_ICONV
int str_iconv(
    const char *fromcode,       /* charset of source string*/
    const char *tocode,     /* charset of destination string */
    char *src,      /* source string */
    int srclen,         /* source string length */
    char *dst,          /* destination string */
    int dstlen)         /* destination string length */
/* 這個函式會將一個字串 (src) 從 charset=fromcode 轉成 charset=tocode,
   srclen 是 src 的長度, dst 是輸出的buffer, dstlen 則指定了
   dst 的大小, 最後會補 '\0', 所以要留一個byte給'\0'.
   如果遇到 src 中有非字集的字, 或是 src 中有未完整的 byte,
   都會砍掉.
 */
{
    iconv_t iconv_descriptor;
    int iconv_ret,dstlen_old;

    dstlen--;   /* keep space for '\0' */

    dstlen_old = dstlen;

    /* Open a descriptor for iconv */
    iconv_descriptor = iconv_open(tocode, fromcode);

    if (iconv_descriptor == ((iconv_t)(-1)) ) /* if open fail */
    {
    strncpy(dst,src,dstlen);
    return dstlen;
    }

    /* Start translation */
    while (srclen > 0 && dstlen > 0)
    {
    iconv_ret = iconv(iconv_descriptor, &src, &srclen,
        &dst, &dstlen);
    if (iconv_ret  != 0)
    {
        switch(errno)
        {
        /* invalid multibyte happened */
        case EILSEQ:
            /* forward that byte */
            *dst = *src;
            src++; srclen--;
            dst++; dstlen--;
            break;
            /* incomplete multibyte happened */
        case EINVAL:
            /* forward that byte (maybe wrong)*/
            *dst = *src;
            src++; srclen--;
            dst++; dstlen--;
            break;
            /* dst no rooms */
        case E2BIG:
            /* break out the while loop */
            srclen = 0;
            break;
        }
    }
    }
    *dst = '\0';
    /* close descriptor of iconv*/
    iconv_close(iconv_descriptor);

    return (dstlen_old - dstlen);
}
#endif


void str_decode_M3(unsigned char *str)
{
    int adj;
    int i;
    unsigned char *src, *dst;
    unsigned char buf[512];
    unsigned char charset[512],dst1[512];


    src = str;
    dst = buf;
    adj = 0;

    while (*src && (dst - buf) < sizeof(buf) - 1)
    {
    if (*src != '=')
    {                        /* Thor: not coded */
        unsigned char *tmp = src;
        while(adj && *tmp && isspace(*tmp)) tmp++;
        if(adj && *tmp=='=')
        {                     /* Thor: jump over space */
        adj = 0;
        src = tmp;
        }
        else
        *dst++ = *src++;
        /* continue;*/        /* Thor: take out */
    }
    else                    /* Thor: *src == '=' */
    {
        unsigned char *tmp = src + 1;
        if(*tmp == '?')       /* Thor: =? coded */
        {
        /* "=?%s?Q?" for QP, "=?%s?B?" for BASE64 */
        tmp ++;
        i=0;
        while(*tmp && *tmp != '?')
        {
            if (i+1<sizeof(charset))
            {
            charset[i] = *tmp;
            charset[i+1]='\0';
            i++;
            }
            tmp++;
        }
        if(*tmp && tmp[1] && tmp[2]=='?') /* Thor: *tmp == '?' */
        {
#ifdef USE_ICONV
            int i = mmdecode(tmp + 3, tmp[1], dst1);
            i = str_iconv(charset,"big5",dst1,i,dst,
                sizeof(buf)-((int)(dst-buf)) );
#else
            int i = mmdecode(tmp + 3, tmp[1], dst);
#endif
            if (i >= 0)
            {
            tmp += 3;       /* Thor: decode's src */
#if 0
            while(*tmp++ != '?'); /* Thor: no ? end, mmdecode -1 */
#endif
            while(*tmp && *tmp++ != '?'); /* Thor: no ? end, mmdecode -1 */
            /* Thor.980901: 0 也算 decode 結束 */
            if(*tmp == '=') tmp++;
            src = tmp;      /* Thor: decode over */
            dst += i;
            adj = 1;        /* Thor: adjcent */
            }
        }
        }

        while(src != tmp)     /* Thor: not coded */
        *dst++ = *src++;
    }
    }
    *dst = 0;
    strcpy(str, buf);
}