util.c

AP_DECLARE(apr_status_t) ap_cfg_getc(char *ch, ap_configfile_t *cfp)
{
    apr_status_t rc = cfp->getch(ch, cfp->param);
    if (rc == APR_SUCCESS && *ch == LF)
        ++cfp->line_number;
    return rc;
}

AP_DECLARE(const char *) ap_pcfg_strerror(apr_pool_t *p, ap_configfile_t *cfp,
                                          apr_status_t rc)
{
    if (rc == APR_SUCCESS)
        return NULL;

    if (rc == APR_ENOSPC)
        return apr_psprintf(p, "Error reading %s at line %d: Line too long",
                            cfp->name, cfp->line_number);

    return apr_psprintf(p, "Error reading %s at line %d: %pm",
                        cfp->name, cfp->line_number, &rc);
}

/* Read one line from open ap_configfile_t, strip LF, increase line number */
/* If custom handler does not define a getstr() function, read char by char */
static apr_status_t ap_cfg_getline_core(char *buf, apr_size_t bufsize,
                                        apr_size_t offset, ap_configfile_t *cfp)
{
    apr_status_t rc;
    /* If a "get string" function is defined, use it */
    if (cfp->getstr != NULL) {
        char *cp;
        char *cbuf = buf + offset;
        apr_size_t cbufsize = bufsize - offset;

        while (1) {
            ++cfp->line_number;
            rc = cfp->getstr(cbuf, cbufsize, cfp->param);
            if (rc == APR_EOF) {
                if (cbuf != buf + offset) {
                    *cbuf = '\0';
                    break;
                }
                else {
                    return APR_EOF;
                }
            }
            if (rc != APR_SUCCESS) {
                return rc;
            }

            /*
             *  check for line continuation,
             *  i.e. match [^\\]\\[\r]\n only
             */
            cp = cbuf;
            cp += strlen(cp);
            if (cp > buf && cp[-1] == LF) {
                cp--;
                if (cp > buf && cp[-1] == CR)
                    cp--;
                if (cp > buf && cp[-1] == '\\') {
                    cp--;
                    /*
                     * line continuation requested -
                     * then remove backslash and continue
                     */
                    cbufsize -= (cp-cbuf);
                    cbuf = cp;
                    continue;
                }
            }
            else if (cp - buf >= bufsize - 1) {
                return APR_ENOSPC;
            }
            break;
        }
    } else {
        /* No "get string" function defined; read character by character */
        apr_size_t i = offset;

        if (bufsize < 2) {
            /* too small, assume caller is crazy */
            return APR_EINVAL;
        }
        buf[offset] = '\0';

        while (1) {
            char c;
            rc = cfp->getch(&c, cfp->param);
            if (rc == APR_EOF) {
                if (i > offset)
                    break;
                else
                    return APR_EOF;
            }
            if (rc != APR_SUCCESS)
                return rc;
            if (c == LF) {
                ++cfp->line_number;
                /* check for line continuation */
                if (i > 0 && buf[i-1] == '\\') {
                    i--;
                    continue;
                }
                else {
                    break;
                }
            }
            buf[i] = c;
            ++i;
            if (i >= bufsize - 1) {
                return APR_ENOSPC;
            }
        }
        buf[i] = '\0';
    }
    return APR_SUCCESS;
}

static int cfg_trim_line(char *buf)
{
    char *start, *end;
    /*
     * Leading and trailing white space is eliminated completely
     */
    start = buf;
    while (apr_isspace(*start))
        ++start;
    /* blast trailing whitespace */
    end = &start[strlen(start)];
    while (--end >= start && apr_isspace(*end))
        *end = '\0';
    /* Zap leading whitespace by shifting */
    if (start != buf)
        memmove(buf, start, end - start + 2);
#ifdef DEBUG_CFG_LINES
    ap_log_error(APLOG_MARK, APLOG_NOTICE, 0, NULL, APLOGNO(00555) "Read config: '%s'", buf);
#endif
    return end - start + 1;
}

/* Read one line from open ap_configfile_t, strip LF, increase line number */
/* If custom handler does not define a getstr() function, read char by char */
AP_DECLARE(apr_status_t) ap_cfg_getline(char *buf, apr_size_t bufsize,
                                        ap_configfile_t *cfp)
{
    apr_status_t rc = ap_cfg_getline_core(buf, bufsize, 0, cfp);
    if (rc == APR_SUCCESS)
        cfg_trim_line(buf);
    return rc;
}

AP_DECLARE(apr_status_t) ap_varbuf_cfg_getline(struct ap_varbuf *vb,
                                               ap_configfile_t *cfp,
                                               apr_size_t max_len)
{
    apr_status_t rc;
    apr_size_t new_len;
    vb->strlen = 0;
    *vb->buf = '\0';

    if (vb->strlen == AP_VARBUF_UNKNOWN)
        vb->strlen = strlen(vb->buf);
    if (vb->avail - vb->strlen < 3) {
        new_len = vb->avail * 2;
        if (new_len > max_len)
            new_len = max_len;
        else if (new_len < 3)
            new_len = 3;
        ap_varbuf_grow(vb, new_len);
    }

    for (;;) {
        rc = ap_cfg_getline_core(vb->buf, vb->avail, vb->strlen, cfp);
        if (rc == APR_ENOSPC || rc == APR_SUCCESS)
            vb->strlen += strlen(vb->buf + vb->strlen);
        if (rc != APR_ENOSPC)
            break;
        if (vb->avail >= max_len)
            return APR_ENOSPC;
        new_len = vb->avail * 2;
        if (new_len > max_len)
            new_len = max_len;
        ap_varbuf_grow(vb, new_len);
        --cfp->line_number;
    }
    if (vb->strlen > max_len)
        return APR_ENOSPC;
    if (rc == APR_SUCCESS)
        vb->strlen = cfg_trim_line(vb->buf);
    return rc;
}

/* Size an HTTP header field list item, as separated by a comma.
 * The return value is a pointer to the beginning of the non-empty list item
 * within the original string (or NULL if there is none) and the address
 * of field is shifted to the next non-comma, non-whitespace character.
 * len is the length of the item excluding any beginning whitespace.
 */
AP_DECLARE(const char *) ap_size_list_item(const char **field, int *len)
{
    const unsigned char *ptr = (const unsigned char *)*field;
    const unsigned char *token;
    int in_qpair, in_qstr, in_com;

    /* Find first non-comma, non-whitespace byte */

    while (*ptr == ',' || apr_isspace(*ptr))
        ++ptr;

    token = ptr;

    /* Find the end of this item, skipping over dead bits */

    for (in_qpair = in_qstr = in_com = 0;
         *ptr && (in_qpair || in_qstr || in_com || *ptr != ',');
         ++ptr) {

        if (in_qpair) {
            in_qpair = 0;
        }
        else {
            switch (*ptr) {
                case '\\': in_qpair = 1;      /* quoted-pair         */
                           break;
                case '"' : if (!in_com)       /* quoted string delim */
                               in_qstr = !in_qstr;
                           break;
                case '(' : if (!in_qstr)      /* comment (may nest)  */
                               ++in_com;
                           break;
                case ')' : if (in_com)        /* end comment         */
                               --in_com;
                           break;
                default  : break;
            }
        }
    }

    if ((*len = (ptr - token)) == 0) {
        *field = (const char *)ptr;
        return NULL;
    }

    /* Advance field pointer to the next non-comma, non-white byte */

    while (*ptr == ',' || apr_isspace(*ptr))
        ++ptr;

    *field = (const char *)ptr;
    return (const char *)token;
}

/* Retrieve an HTTP header field list item, as separated by a comma,
 * while stripping insignificant whitespace and lowercasing anything not in
 * a quoted string or comment.  The return value is a new string containing
 * the converted list item (or NULL if none) and the address pointed to by
 * field is shifted to the next non-comma, non-whitespace.
 */
AP_DECLARE(char *) ap_get_list_item(apr_pool_t *p, const char **field)
{
    const char *tok_start;
    const unsigned char *ptr;
    unsigned char *pos;
    char *token;
    int addspace = 0, in_qpair = 0, in_qstr = 0, in_com = 0, tok_len = 0;

    /* Find the beginning and maximum length of the list item so that
     * we can allocate a buffer for the new string and reset the field.
     */
    if ((tok_start = ap_size_list_item(field, &tok_len)) == NULL) {
        return NULL;
    }
    token = apr_palloc(p, tok_len + 1);

    /* Scan the token again, but this time copy only the good bytes.
     * We skip extra whitespace and any whitespace around a '=', '/',
     * or ';' and lowercase normal characters not within a comment,
     * quoted-string or quoted-pair.
     */
    for (ptr = (const unsigned char *)tok_start, pos = (unsigned char *)token;
         *ptr && (in_qpair || in_qstr || in_com || *ptr != ',');
         ++ptr) {

        if (in_qpair) {
            in_qpair = 0;
            *pos++ = *ptr;
        }
        else {
            switch (*ptr) {
                case '\\': in_qpair = 1;
                           if (addspace == 1)
                               *pos++ = ' ';
                           *pos++ = *ptr;
                           addspace = 0;
                           break;
                case '"' : if (!in_com)
                               in_qstr = !in_qstr;
                           if (addspace == 1)
                               *pos++ = ' ';
                           *pos++ = *ptr;
                           addspace = 0;
                           break;
                case '(' : if (!in_qstr)
                               ++in_com;
                           if (addspace == 1)
                               *pos++ = ' ';
                           *pos++ = *ptr;
                           addspace = 0;
                           break;
                case ')' : if (in_com)
                               --in_com;
                           *pos++ = *ptr;
                           addspace = 0;
                           break;
                case ' ' :
                case '\t': if (addspace)
                               break;
                           if (in_com || in_qstr)
                               *pos++ = *ptr;
                           else
                               addspace = 1;
                           break;
                case '=' :
                case '/' :
                case ';' : if (!(in_com || in_qstr))
                               addspace = -1;
                           *pos++ = *ptr;
                           break;
                default  : if (addspace == 1)
                               *pos++ = ' ';
                           *pos++ = (in_com || in_qstr) ? *ptr
                                                        : apr_tolower(*ptr);
                           addspace = 0;
                           break;
            }
        }
    }
    *pos = '\0';

    return token;
}

typedef enum ap_etag_e {
    AP_ETAG_NONE,
    AP_ETAG_WEAK,
    AP_ETAG_STRONG
} ap_etag_e;

/* Find an item in canonical form (lowercase, no extra spaces) within
 * an HTTP field value list.  Returns 1 if found, 0 if not found.
 * This would be much more efficient if we stored header fields as
 * an array of list items as they are received instead of a plain string.
 */
static int find_list_item(apr_pool_t *p, const char *line,
                                  const char *tok, ap_etag_e type)
{
    const unsigned char *pos;
    const unsigned char *ptr = (const unsigned char *)line;
    int good = 0, addspace = 0, in_qpair = 0, in_qstr = 0, in_com = 0;

    if (!line || !tok) {
        return 0;
    }
    if (type == AP_ETAG_STRONG && *tok != '\"') {
        return 0;
    }
    if (type == AP_ETAG_WEAK) {
        if (*tok == 'W' && (*(tok+1)) == '/' && (*(tok+2)) == '\"') {
            tok += 2;
        }
        else if (*tok != '\"') {
            return 0;
        }
    }

    do {  /* loop for each item in line's list */

        /* Find first non-comma, non-whitespace byte */
        while (*ptr == ',' || apr_isspace(*ptr)) {
            ++ptr;
        }

        /* Account for strong or weak Etags, depending on our search */
        if (type == AP_ETAG_STRONG && *ptr != '\"') {
            break;
        }
        if (type == AP_ETAG_WEAK) {
            if (*ptr == 'W' && (*(ptr+1)) == '/' && (*(ptr+2)) == '\"') {
                ptr += 2;
            }
            else if (*ptr != '\"') {
                break;
            }
        }

        if (*ptr)
            good = 1;  /* until proven otherwise for this item */
        else
            break;     /* no items left and nothing good found */

        /* We skip extra whitespace and any whitespace around a '=', '/',
         * or ';' and lowercase normal characters not within a comment,
         * quoted-string or quoted-pair.
         */
        for (pos = (const unsigned char *)tok;
             *ptr && (in_qpair || in_qstr || in_com || *ptr != ',');
             ++ptr) {

            if (in_qpair) {
                in_qpair = 0;
                if (good)
                    good = (*pos++ == *ptr);
            }
            else {
                switch (*ptr) {
                    case '\\': in_qpair = 1;
                               if (addspace == 1)
                                   good = good && (*pos++ == ' ');
                               good = good && (*pos++ == *ptr);
                               addspace = 0;
                               break;
                    case '"' : if (!in_com)
                                   in_qstr = !in_qstr;
                               if (addspace == 1)
                                   good = good && (*pos++ == ' ');
                               good = good && (*pos++ == *ptr);
                               addspace = 0;
                               break;
                    case '(' : if (!in_qstr)
                                   ++in_com;
                               if (addspace == 1)
                                   good = good && (*pos++ == ' ');
                               good = good && (*pos++ == *ptr);
                               addspace = 0;
                               break;
                    case ')' : if (in_com)
                                   --in_com;
                               good = good && (*pos++ == *ptr);
                               addspace = 0;
                               break;
                    case ' ' :
                    case '\t': if (addspace || !good)
                                   break;
                               if (in_com || in_qstr)
                                   good = (*pos++ == *ptr);
                               else
                                   addspace = 1;
                               break;
                    case '=' :
                    case '/' :
                    case ';' : if (!(in_com || in_qstr))
                                   addspace = -1;
                               good = good && (*pos++ == *ptr);
                               break;
                    default  : if (!good)
                                   break;
                               if (addspace == 1)
                                   good = (*pos++ == ' ');
                               if (in_com || in_qstr)
                                   good = good && (*pos++ == *ptr);
                               else
                                   good = good
                                       && (apr_tolower(*pos++) == apr_tolower(*ptr));
                               addspace = 0;
                               break;
                }
            }
        }
        if (good && *pos)
            good = 0;          /* not good if only a prefix was matched */

    } while (*ptr && !good);

    return good;
}

/* Find an item in canonical form (lowercase, no extra spaces) within
 * an HTTP field value list.  Returns 1 if found, 0 if not found.
 * This would be much more efficient if we stored header fields as
 * an array of list items as they are received instead of a plain string.
 */
AP_DECLARE(int) ap_find_list_item(apr_pool_t *p, const char *line,
                                  const char *tok)
{
    return find_list_item(p, line, tok, AP_ETAG_NONE);
}

/* Find a strong Etag in canonical form (lowercase, no extra spaces) within
 * an HTTP field value list.  Returns 1 if found, 0 if not found.
 */
AP_DECLARE(int) ap_find_etag_strong(apr_pool_t *p, const char *line,
                                    const char *tok)
{
    return find_list_item(p, line, tok, AP_ETAG_STRONG);
}

/* Find a weak ETag in canonical form (lowercase, no extra spaces) within
 * an HTTP field value list.  Returns 1 if found, 0 if not found.
 */
AP_DECLARE(int) ap_find_etag_weak(apr_pool_t *p, const char *line,
                                  const char *tok)
{
    return find_list_item(p, line, tok, AP_ETAG_WEAK);
}

/* Grab a list of tokens of the format 1#token (from RFC7230) */
AP_DECLARE(const char *) ap_parse_token_list_strict(apr_pool_t *p,
                                                const char *str_in,
                                                apr_array_header_t **tokens,
                                                int skip_invalid)
{
    int in_leading_space = 1;
    int in_trailing_space = 0;
    int string_end = 0;
    const char *tok_begin;
    const char *cur;

    if (!str_in) {
        return NULL;
    }

    tok_begin = cur = str_in;

    while (!string_end) {
        const unsigned char c = (unsigned char)*cur;

        if (!TEST_CHAR(c, T_HTTP_TOKEN_STOP)) {
            /* Non-separator character; we are finished with leading
             * whitespace. We must never have encountered any trailing
             * whitespace before the delimiter (comma) */
            in_leading_space = 0;
            if (in_trailing_space) {
                return "Encountered illegal whitespace in token";
            }
        }
        else if (c == ' ' || c == '\t') {
            /* "Linear whitespace" only includes ASCII CRLF, space, and tab;
             * we can't get a CRLF since headers are split on them already,
             * so only look for a space or a tab */
            if (in_leading_space) {
                /* We're still in leading whitespace */
                ++tok_begin;
            }
            else {
                /* We must be in trailing whitespace */
                ++in_trailing_space;
            }
        }
        else if (c == ',' || c == '\0') {
            if (!in_leading_space) {
                /* If we're out of the leading space, we know we've read some
                 * characters of a token */
                if (*tokens == NULL) {
                    *tokens = apr_array_make(p, 4, sizeof(char *));
                }
                APR_ARRAY_PUSH(*tokens, char *) =
                    apr_pstrmemdup((*tokens)->pool, tok_begin,
                                   (cur - tok_begin) - in_trailing_space);
            }
            /* We're allowed to have null elements, just don't add them to the
             * array */

            tok_begin = cur + 1;
            in_leading_space = 1;
            in_trailing_space = 0;
            string_end = (c == '\0');
        }
        else {
            /* Encountered illegal separator char */
            if (skip_invalid) {
                /* Skip to the next separator */
                const char *temp;
                temp = ap_strchr_c(cur, ',');
                if(!temp) {
                    temp = ap_strchr_c(cur, '\0');
                }

                /* Act like we haven't seen a token so we reset */
                cur = temp - 1;
                in_leading_space = 1;
                in_trailing_space = 0;
            }
            else {
                return apr_psprintf(p, "Encountered illegal separator "
                                    "'\\x%.2x'", (unsigned int)c);
            }
        }

        ++cur;
    }

    return NULL;
}

/* Scan a string for HTTP VCHAR/obs-text characters including HT and SP
 * (as used in header values, for example, in RFC 7230 section 3.2)
 * returning the pointer to the first non-HT ASCII ctrl character.
 */
AP_DECLARE(const char *) ap_scan_http_field_content(const char *ptr)
{
    for ( ; !TEST_CHAR(*ptr, T_HTTP_CTRLS); ++ptr) ;

    return ptr;
}

/* Scan a string for HTTP token characters, returning the pointer to
 * the first non-token character.
 */
AP_DECLARE(const char *) ap_scan_http_token(const char *ptr)
{
    for ( ; !TEST_CHAR(*ptr, T_HTTP_TOKEN_STOP); ++ptr) ;

    return ptr;
}

/* Scan a string for visible ASCII (0x21-0x7E) or obstext (0x80+)
 * and return a pointer to the first ctrl/space character encountered.
 */
AP_DECLARE(const char *) ap_scan_vchar_obstext(const char *ptr)
{
    for ( ; TEST_CHAR(*ptr, T_VCHAR_OBSTEXT); ++ptr) ;

    return ptr;
}

/* Retrieve a token, spacing over it and returning a pointer to
 * the first non-white byte afterwards.  Note that these tokens
 * are delimited by semis and commas; and can also be delimited
 * by whitespace at the caller's option.
 */

AP_DECLARE(char *) ap_get_token(apr_pool_t *p, const char **accept_line,
                                int accept_white)
{
    const char *ptr = *accept_line;
    const char *tok_start;
    char *token;

    /* Find first non-white byte */

    while (apr_isspace(*ptr))
        ++ptr;

    tok_start = ptr;

    /* find token end, skipping over quoted strings.
     * (comments are already gone).
     */

    while (*ptr && (accept_white || !apr_isspace(*ptr))
           && *ptr != ';' && *ptr != ',') {
        if (*ptr++ == '"')
            while (*ptr)
                if (*ptr++ == '"')
                    break;
    }

    token = apr_pstrmemdup(p, tok_start, ptr - tok_start);

    /* Advance accept_line pointer to the next non-white byte */

    while (apr_isspace(*ptr))
        ++ptr;

    *accept_line = ptr;
    return token;
}


/* find http tokens, see the definition of token from RFC2068 */
AP_DECLARE(int) ap_find_token(apr_pool_t *p, const char *line, const char *tok)
{
    const unsigned char *start_token;
    const unsigned char *s;

    if (!line)
        return 0;

    s = (const unsigned char *)line;
    for (;;) {
        /* find start of token, skip all stop characters */
        while (*s && TEST_CHAR(*s, T_HTTP_TOKEN_STOP)) {
            ++s;
        }
        if (!*s) {
            return 0;
        }
        start_token = s;
        /* find end of the token */
        while (*s && !TEST_CHAR(*s, T_HTTP_TOKEN_STOP)) {
            ++s;
        }
        if (!strncasecmp((const char *)start_token, (const char *)tok,
                         s - start_token)) {
            return 1;
        }
        if (!*s) {
            return 0;
        }
    }
}


AP_DECLARE(int) ap_find_last_token(apr_pool_t *p, const char *line,
                                   const char *tok)
{
    int llen, tlen, lidx;

    if (!line)
        return 0;

    llen = strlen(line);
    tlen = strlen(tok);
    lidx = llen - tlen;

    if (lidx < 0 ||
        (lidx > 0 && !(apr_isspace(line[lidx - 1]) || line[lidx - 1] == ',')))
        return 0;

    return (strncasecmp(&line[lidx], tok, tlen) == 0);
}

AP_DECLARE(char *) ap_escape_shell_cmd(apr_pool_t *p, const char *str)
{
    char *cmd;
    unsigned char *d;
    const unsigned char *s;

    cmd = apr_palloc(p, 2 * strlen(str) + 1);        /* Be safe */
    d = (unsigned char *)cmd;
    s = (const unsigned char *)str;
    for (; *s; ++s) {

#if defined(OS2) || defined(WIN32)
        /*
         * Newlines to Win32/OS2 CreateProcess() are ill advised.
         * Convert them to spaces since they are effectively white
         * space to most applications
         */
        if (*s == '\r' || *s == '\n') {
             *d++ = ' ';
             continue;
         }
#endif

        if (TEST_CHAR(*s, T_ESCAPE_SHELL_CMD)) {
            *d++ = '\\';
        }
        *d++ = *s;
    }
    *d = '\0';

    return cmd;
}

static char x2c(const char *what)
{
    char digit;

#if !APR_CHARSET_EBCDIC
    digit = ((what[0] >= 'A') ? ((what[0] & 0xdf) - 'A') + 10
             : (what[0] - '0'));
    digit *= 16;
    digit += (what[1] >= 'A' ? ((what[1] & 0xdf) - 'A') + 10
              : (what[1] - '0'));
#else /*APR_CHARSET_EBCDIC*/
    char xstr[5];
    xstr[0]='0';
    xstr[1]='x';
    xstr[2]=what[0];
    xstr[3]=what[1];
    xstr[4]='\0';
    digit = apr_xlate_conv_byte(ap_hdrs_from_ascii,
                                0xFF & strtol(xstr, NULL, 16));
#endif /*APR_CHARSET_EBCDIC*/
    return (digit);
}

/*
 * Unescapes a URL, leaving reserved characters intact.
 * Returns 0 on success, non-zero on error
 * Failure is due to
 *   bad % escape       returns HTTP_BAD_REQUEST
 *
 *   decoding %00 or a forbidden character returns HTTP_NOT_FOUND
 */

static int unescape_url(char *url, const char *forbid, const char *reserved)
{
    int badesc, badpath;
    char *x, *y;

    badesc = 0;
    badpath = 0;
    /* Initial scan for first '%'. Don't bother writing values before
     * seeing a '%' */
    y = strchr(url, '%');
    if (y == NULL) {
        return OK;
    }
    for (x = y; *y; ++x, ++y) {
        if (*y != '%') {
            *x = *y;
        }
        else {
            if (!apr_isxdigit(*(y + 1)) || !apr_isxdigit(*(y + 2))) {
                badesc = 1;
                *x = '%';
            }
            else {
                char decoded;
                decoded = x2c(y + 1);
                if ((decoded == '\0')
                    || (forbid && ap_strchr_c(forbid, decoded))) {
                    badpath = 1;
                    *x = decoded;
                    y += 2;
                }
                else if (reserved && ap_strchr_c(reserved, decoded)) {
                    *x++ = *y++;
                    *x++ = *y++;
                    *x = *y;
                }
                else {
                    *x = decoded;
                    y += 2;
                }
            }
        }
    }
    *x = '\0';
    if (badesc) {
        return HTTP_BAD_REQUEST;
    }
    else if (badpath) {
        return HTTP_NOT_FOUND;
    }
    else {
        return OK;
    }
}
AP_DECLARE(int) ap_unescape_url(char *url)
{
    /* Traditional */
    return unescape_url(url, SLASHES, NULL);
}
AP_DECLARE(int) ap_unescape_url_keep2f(char *url, int decode_slashes)
{
    /* AllowEncodedSlashes (corrected) */
    if (decode_slashes) {
        /* no chars reserved */
        return unescape_url(url, NULL, NULL);
    } else {
        /* reserve (do not decode) encoded slashes */
        return unescape_url(url, NULL, SLASHES);
    }
}
#ifdef NEW_APIS
/* IFDEF these out until they've been thought through.
 * Just a germ of an API extension for now
 */
AP_DECLARE(int) ap_unescape_url_proxy(char *url)
{
    /* leave RFC1738 reserved characters intact, * so proxied URLs
     * don't get mangled.  Where does that leave encoded '&' ?
     */
    return unescape_url(url, NULL, "/;?");
}
AP_DECLARE(int) ap_unescape_url_reserved(char *url, const char *reserved)
{
    return unescape_url(url, NULL, reserved);
}
#endif

AP_DECLARE(int) ap_unescape_urlencoded(char *query)
{
    char *slider;

    /* replace plus with a space */
    if (query) {
        for (slider = query; *slider; slider++) {
            if (*slider == '+') {
                *slider = ' ';
            }
        }
    }

    /* unescape everything else */
    return unescape_url(query, NULL, NULL);
}

AP_DECLARE(char *) ap_construct_server(apr_pool_t *p, const char *hostname,
                                       apr_port_t port, const request_rec *r)
{
    if (ap_is_default_port(port, r)) {
        return apr_pstrdup(p, hostname);
    }
    else {
        return apr_psprintf(p, "%s:%u", hostname, port);
    }
}

AP_DECLARE(int) ap_unescape_all(char *url)
{
    return unescape_url(url, NULL, NULL);
}

/* c2x takes an unsigned, and expects the caller has guaranteed that
 * 0 <= what < 256... which usually means that you have to cast to
 * unsigned char first, because (unsigned)(char)(x) first goes through
 * signed extension to an int before the unsigned cast.
 *
 * The reason for this assumption is to assist gcc code generation --
 * the unsigned char -> unsigned extension is already done earlier in
 * both uses of this code, so there's no need to waste time doing it
 * again.
 */
static const char c2x_table[] = "0123456789abcdef";

static APR_INLINE unsigned char *c2x(unsigned what, unsigned char prefix,
                                     unsigned char *where)
{
#if APR_CHARSET_EBCDIC
    what = apr_xlate_conv_byte(ap_hdrs_to_ascii, (unsigned char)what);
#endif /*APR_CHARSET_EBCDIC*/
    *where++ = prefix;
    *where++ = c2x_table[what >> 4];
    *where++ = c2x_table[what & 0xf];
    return where;
}

/*
 * escape_path_segment() escapes a path segment, as defined in RFC 1808. This
 * routine is (should be) OS independent.
 *
 * os_escape_path() converts an OS path to a URL, in an OS dependent way. In all
 * cases if a ':' occurs before the first '/' in the URL, the URL should be
 * prefixed with "./" (or the ':' escaped). In the case of Unix, this means
 * leaving '/' alone, but otherwise doing what escape_path_segment() does. For
 * efficiency reasons, we don't use escape_path_segment(), which is provided for
 * reference. Again, RFC 1808 is where this stuff is defined.
 *
 * If partial is set, os_escape_path() assumes that the path will be appended to
 * something with a '/' in it (and thus does not prefix "./").
 */

AP_DECLARE(char *) ap_escape_path_segment_buffer(char *copy, const char *segment)
{
    const unsigned char *s = (const unsigned char *)segment;
    unsigned char *d = (unsigned char *)copy;
    unsigned c;

    while ((c = *s)) {
        if (TEST_CHAR(c, T_ESCAPE_PATH_SEGMENT)) {
            d = c2x(c, '%', d);
        }
        else {
            *d++ = c;
        }
        ++s;
    }
    *d = '\0';
    return copy;
}

AP_DECLARE(char *) ap_escape_path_segment(apr_pool_t *p, const char *segment)
{
    return ap_escape_path_segment_buffer(apr_palloc(p, 3 * strlen(segment) + 1), segment);
}

AP_DECLARE(char *) ap_os_escape_path(apr_pool_t *p, const char *path, int partial)
{
    char *copy = apr_palloc(p, 3 * strlen(path) + 3);
    const unsigned char *s = (const unsigned char *)path;
    unsigned char *d = (unsigned char *)copy;
    unsigned c;

    if (!partial) {
        const char *colon = ap_strchr_c(path, ':');
        const char *slash = ap_strchr_c(path, '/');

        if (colon && (!slash || colon < slash)) {
            *d++ = '.';
            *d++ = '/';
        }
    }
    while ((c = *s)) {
        if (TEST_CHAR(c, T_OS_ESCAPE_PATH)) {
            d = c2x(c, '%', d);
        }
        else {
            *d++ = c;
        }
        ++s;
    }
    *d = '\0';
    return copy;
}