jszip-esm/lib/utf8.js

import * as utils from './utils';
import support from './support';
import GenericWorker from './stream/GenericWorker';

/* eslint-disable */

/**
 * The following functions come from pako, from pako/lib/utils/strings
 * released under the MIT license, see pako https://github.com/nodeca/pako/
 */

// Returns the utf8 lengths (calculated by first byte of sequence)
// Note, that 5 & 6-byte values and some 4-byte values can not be represented in JS,
// because max possible codepoint is 0x10ffff
let utf8len = function(c) {
    const _utf8len = new Array(256);
    for (let i = 0; i < 256; i++) {
        _utf8len[i] = (i >= 252 ? 6 : i >= 248 ? 5 : i >= 240 ? 4 : i >= 224 ? 3 : i >= 192 ? 2 : 1);
    }
    _utf8len[254] = _utf8len[254] = 1; // Invalid sequence start

    // Memoize table after first call
    utf8len = function(c) {
        return _utf8len[c];
    };

    return _utf8len[c];
};

// convert string to array (typed, when possible)
var string2buf = function (str) {
    var buf, c, c2, m_pos, i, str_len = str.length, buf_len = 0;

    // count binary size
    for (m_pos = 0; m_pos < str_len; m_pos++) {
        c = str.charCodeAt(m_pos);
        if ((c & 0xfc00) === 0xd800 && (m_pos+1 < str_len)) {
            c2 = str.charCodeAt(m_pos+1);
            if ((c2 & 0xfc00) === 0xdc00) {
                c = 0x10000 + ((c - 0xd800) << 10) + (c2 - 0xdc00);
                m_pos++;
            }
        }
        buf_len += c < 0x80 ? 1 : c < 0x800 ? 2 : c < 0x10000 ? 3 : 4;
    }

    // allocate buffer
    if (support.uint8array) {
        buf = new Uint8Array(buf_len);
    } else {
        buf = new Array(buf_len);
    }

    // convert
    for (i=0, m_pos = 0; i < buf_len; m_pos++) {
        c = str.charCodeAt(m_pos);
        if ((c & 0xfc00) === 0xd800 && (m_pos+1 < str_len)) {
            c2 = str.charCodeAt(m_pos+1);
            if ((c2 & 0xfc00) === 0xdc00) {
                c = 0x10000 + ((c - 0xd800) << 10) + (c2 - 0xdc00);
                m_pos++;
            }
        }
        if (c < 0x80) {
            /* one byte */
            buf[i++] = c;
        } else if (c < 0x800) {
            /* two bytes */
            buf[i++] = 0xC0 | (c >>> 6);
            buf[i++] = 0x80 | (c & 0x3f);
        } else if (c < 0x10000) {
            /* three bytes */
            buf[i++] = 0xE0 | (c >>> 12);
            buf[i++] = 0x80 | (c >>> 6 & 0x3f);
            buf[i++] = 0x80 | (c & 0x3f);
        } else {
            /* four bytes */
            buf[i++] = 0xf0 | (c >>> 18);
            buf[i++] = 0x80 | (c >>> 12 & 0x3f);
            buf[i++] = 0x80 | (c >>> 6 & 0x3f);
            buf[i++] = 0x80 | (c & 0x3f);
        }
    }

    return buf;
};

// Calculate max possible position in utf8 buffer,
// that will not break sequence. If that's not possible
// - (very small limits) return max size as is.
//
// buf[] - utf8 bytes array
// max   - length limit (mandatory);
var utf8border = function(buf, max) {
    var pos;

    max = max || buf.length;
    if (max > buf.length) { max = buf.length; }

    // go back from last position, until start of sequence found
    pos = max-1;
    while (pos >= 0 && (buf[pos] & 0xC0) === 0x80) { pos--; }

    // Fuckup - very small and broken sequence,
    // return max, because we should return something anyway.
    if (pos < 0) { return max; }

    // If we came to start of buffer - that means vuffer is too small,
    // return max too.
    if (pos === 0) { return max; }

    return (pos + utf8len(buf[pos]) > max) ? pos : max;
};

// convert array to string
var buf2string = function (buf) {
    var str, i, out, c, c_len;
    var len = buf.length;

    // Reserve max possible length (2 words per char)
    // NB: by unknown reasons, Array is significantly faster for
    //     String.fromCharCode.apply than Uint16Array.
    var utf16buf = new Array(len*2);

    for (out=0, i=0; i<len;) {
        c = buf[i++];
        // quick process ascii
        if (c < 0x80) { utf16buf[out++] = c; continue; }

        c_len = utf8len(c);
        // skip 5 & 6 byte codes
        if (c_len > 4) { utf16buf[out++] = 0xfffd; i += c_len-1; continue; }

        // apply mask on first byte
        c &= c_len === 2 ? 0x1f : c_len === 3 ? 0x0f : 0x07;
        // join the rest
        while (c_len > 1 && i < len) {
            c = (c << 6) | (buf[i++] & 0x3f);
            c_len--;
        }

        // terminated by end of string?
        if (c_len > 1) { utf16buf[out++] = 0xfffd; continue; }

        if (c < 0x10000) {
            utf16buf[out++] = c;
        } else {
            c -= 0x10000;
            utf16buf[out++] = 0xd800 | ((c >> 10) & 0x3ff);
            utf16buf[out++] = 0xdc00 | (c & 0x3ff);
        }
    }

    // shrinkBuf(utf16buf, out)
    if (utf16buf.length !== out) {
        if(utf16buf.subarray) {
            utf16buf = utf16buf.subarray(0, out);
        } else {
            utf16buf.length = out;
        }
    }

    // return String.fromCharCode.apply(null, utf16buf);
    return utils.applyFromCharCode(utf16buf);
};


// That's all for the pako functions.


/**
 * Transform a javascript string into an array (typed if possible) of bytes,
 * UTF-8 encoded.
 * @param {String} str the string to encode
 * @return {Array|Uint8Array|Buffer} the UTF-8 encoded string.
 */
export const utf8encode = function utf8encode(str) {
    return string2buf(str);
};

/**
 * Transform a bytes array (or a representation) representing an UTF-8 encoded
 * string into a javascript string.
 * @param {Array|Uint8Array|Buffer} buf the data de decode
 * @return {String} the decoded string.
 */
export const utf8decode = function utf8decode(buf) {
    buf = utils.transformTo(support.uint8array ? "uint8array" : "array", buf);

    return buf2string(buf);
};

/**
 * A worker to decode utf8 encoded binary chunks into string chunks.
 * @constructor
 */
class Utf8DecodeWorker extends GenericWorker {
    constructor() {
        super("utf-8 decode");

        // the last bytes if a chunk didn't end with a complete codepoint.
        this.leftOver = null;
    }

    /**
     * @see GenericWorker.processChunk
     */
    processChunk (chunk) {
        var data = utils.transformTo(support.uint8array ? "uint8array" : "array", chunk.data);

        // 1st step, re-use what's left of the previous chunk
        if (this.leftOver && this.leftOver.length) {
            if(support.uint8array) {
                var previousData = data;
                data = new Uint8Array(previousData.length + this.leftOver.length);
                data.set(this.leftOver, 0);
                data.set(previousData, this.leftOver.length);
            } else {
                data = this.leftOver.concat(data);
            }
            this.leftOver = null;
        }

        var nextBoundary = utf8border(data);
        var usableData = data;
        if (nextBoundary !== data.length) {
            if (support.uint8array) {
                usableData = data.subarray(0, nextBoundary);
                this.leftOver = data.subarray(nextBoundary, data.length);
            } else {
                usableData = data.slice(0, nextBoundary);
                this.leftOver = data.slice(nextBoundary, data.length);
            }
        }

        this.push({
            data : utf8decode(usableData),
            meta : chunk.meta
        });
    }

    /**
     * @see GenericWorker.flush
     */
    flush() {
        if (this.leftOver && this.leftOver.length) {
            this.push({
                data : utf8decode(this.leftOver),
                meta : {}
            });
            this.leftOver = null;
        }
    }
}

/**
 * A worker to endcode string chunks into utf8 encoded binary chunks.
 * @constructor
 */
class Utf8EncodeWorker extends GenericWorker {
    constructor() {
        super("utf-8 encode");
    }

    /**
     * @see GenericWorker.processChunk
     */
    processChunk(chunk) {
        this.push({
            data: utf8encode(chunk.data),
            meta: chunk.meta
        });
    }
}

export { Utf8DecodeWorker, Utf8EncodeWorker };