zlib.d

// Written in the D programming language.

/**
 * Compress/decompress data using the $(HTTP www.zlib.net, zlib library).
 *
 * Examples:
 *
 * If you have a small buffer you can use $(LREF compress) and
 * $(LREF uncompress) directly.
 *
 * -------
 * import std.zlib;
 *
 * auto src =
 * "the quick brown fox jumps over the lazy dog\r
 *  the quick brown fox jumps over the lazy dog\r";
 *
 * ubyte[] dst;
 * ubyte[] result;
 *
 * dst = compress(src);
 * result = cast(ubyte[]) uncompress(dst);
 * assert(result == src);
 * -------
 *
 * When the data to be compressed doesn't fit in one buffer, use
 * $(LREF Compress) and $(LREF UnCompress).
 *
 * -------
 * import std.zlib;
 * import std.stdio;
 * import std.conv : to;
 * import std.algorithm.iteration : map;
 *
 * UnCompress decmp = new UnCompress;
 * foreach (chunk; stdin.byChunk(4096).map!(x => decmp.uncompress(x)))
 * {
 *     chunk.to!string.write;
 * }

 * -------
 *
 * References:
 *  $(HTTP en.wikipedia.org/wiki/Zlib, Wikipedia)
 *
 * Copyright: Copyright The D Language Foundation 2000 - 2011.
 * License:   $(HTTP www.boost.org/LICENSE_1_0.txt, Boost License 1.0).
 * Authors:   $(HTTP digitalmars.com, Walter Bright)
 * Source:    $(PHOBOSSRC std/zlib.d)
 */
/*          Copyright The D Language Foundation 2000 - 2011.
 * Distributed under the Boost Software License, Version 1.0.
 *    (See accompanying file LICENSE_1_0.txt or copy at
 *          http://www.boost.org/LICENSE_1_0.txt)
 */
module std.zlib;

//debug=zlib;       // uncomment to turn on debugging printf's

import etc.c.zlib;

// Values for 'mode'

enum
{
    Z_NO_FLUSH      = 0,
    Z_SYNC_FLUSH    = 2,
    Z_FULL_FLUSH    = 3,
    Z_FINISH        = 4,
}

/*************************************
 * Errors throw a ZlibException.
 */

class ZlibException : Exception
{
    private static string getmsg(int errnum) nothrow @nogc pure @safe
    {
        string msg;
        switch (errnum)
        {
            case Z_STREAM_END:      msg = "stream end"; break;
            case Z_NEED_DICT:       msg = "need dict"; break;
            case Z_ERRNO:           msg = "errno"; break;
            case Z_STREAM_ERROR:    msg = "stream error"; break;
            case Z_DATA_ERROR:      msg = "data error"; break;
            case Z_MEM_ERROR:       msg = "mem error"; break;
            case Z_BUF_ERROR:       msg = "buf error"; break;
            case Z_VERSION_ERROR:   msg = "version error"; break;
            default:                msg = "unknown error";  break;
        }
        return msg;
    }

    this(int errnum)
    {
        super(getmsg(errnum));
    }
}

/**
 * $(P Compute the Adler-32 checksum of a buffer's worth of data.)
 *
 * Params:
 *     adler = the starting checksum for the computation. Use 1
 *             for a new checksum. Use the output of this function
 *             for a cumulative checksum.
 *     buf = buffer containing input data
 *
 * Returns:
 *     A `uint` checksum for the provided input data and starting checksum
 *
 * See_Also:
 *     $(LINK http://en.wikipedia.org/wiki/Adler-32)
 */

uint adler32(uint adler, const(void)[] buf)
{
    import std.range : chunks;
    foreach (chunk; (cast(ubyte[]) buf).chunks(0xFFFF0000))
    {
        adler = etc.c.zlib.adler32(adler, chunk.ptr, cast(uint) chunk.length);
    }
    return adler;
}

///
@system unittest
{
    static ubyte[] data = [1,2,3,4,5,6,7,8,9,10];

    uint adler = adler32(0u, data);
    assert(adler == 0xdc0037);
}

@system unittest
{
    static string data = "test";

    uint adler = adler32(1, data);
    assert(adler == 0x045d01c1);
}

/**
 * $(P Compute the CRC32 checksum of a buffer's worth of data.)
 *
 * Params:
 *     crc = the starting checksum for the computation. Use 0
 *             for a new checksum. Use the output of this function
 *             for a cumulative checksum.
 *     buf = buffer containing input data
 *
 * Returns:
 *     A `uint` checksum for the provided input data and starting checksum
 *
 * See_Also:
 *     $(LINK http://en.wikipedia.org/wiki/Cyclic_redundancy_check)
 */

uint crc32(uint crc, const(void)[] buf)
{
    import std.range : chunks;
    foreach (chunk; (cast(ubyte[]) buf).chunks(0xFFFF0000))
    {
        crc = etc.c.zlib.crc32(crc, chunk.ptr, cast(uint) chunk.length);
    }
    return crc;
}

@system unittest
{
    static ubyte[] data = [1,2,3,4,5,6,7,8,9,10];

    uint crc;

    debug(zlib) printf("D.zlib.crc32.unittest\n");
    crc = crc32(0u, cast(void[]) data);
    debug(zlib) printf("crc = %x\n", crc);
    assert(crc == 0x2520577b);
}

/**
 * $(P Compress data)
 *
 * Params:
 *     srcbuf = buffer containing the data to compress
 *     level = compression level. Legal values are -1 .. 9, with -1 indicating
 *             the default level (6), 0 indicating no compression, 1 being the
 *             least compression and 9 being the most.
 *
 * Returns:
 *     the compressed data
 */

ubyte[] compress(const(void)[] srcbuf, int level)
in
{
    assert(-1 <= level && level <= 9, "Compression level needs to be within [-1, 9].");
}
do
{
    import core.memory : GC;
    import std.array : uninitializedArray;
    auto destlen = srcbuf.length + ((srcbuf.length + 1023) / 1024) + 12;
    auto destbuf = uninitializedArray!(ubyte[])(destlen);
    auto err = etc.c.zlib.compress2(destbuf.ptr, &destlen, cast(ubyte *) srcbuf.ptr, srcbuf.length, level);
    if (err)
    {
        GC.free(destbuf.ptr);
        throw new ZlibException(err);
    }

    destbuf.length = destlen;
    return destbuf;
}

/*********************************************
 * ditto
 */

ubyte[] compress(const(void)[] srcbuf)
{
    return compress(srcbuf, Z_DEFAULT_COMPRESSION);
}

/*********************************************
 * Decompresses the data in srcbuf[].
 * Params:
 *  srcbuf  = buffer containing the compressed data.
 *  destlen = size of the uncompressed data.
 *            It need not be accurate, but the decompression will be faster
 *            if the exact size is supplied.
 *  winbits = the base two logarithm of the maximum window size.
 * Returns: the decompressed data.
 */

void[] uncompress(const(void)[] srcbuf, size_t destlen = 0u, int winbits = 15)
{
    import std.conv : to;
    int err;
    ubyte[] destbuf;

    if (!destlen)
        destlen = srcbuf.length * 2 + 1;

    etc.c.zlib.z_stream zs;
    zs.next_in = cast(typeof(zs.next_in)) srcbuf.ptr;
    zs.avail_in = to!uint(srcbuf.length);
    err = etc.c.zlib.inflateInit2(&zs, winbits);
    if (err)
    {
        throw new ZlibException(err);
    }

    size_t olddestlen = 0u;

    loop:
    while (true)
    {
        destbuf.length = destlen;
        zs.next_out = cast(typeof(zs.next_out)) &destbuf[olddestlen];
        zs.avail_out = to!uint(destlen - olddestlen);
        olddestlen = destlen;

        err = etc.c.zlib.inflate(&zs, Z_NO_FLUSH);
        switch (err)
        {
            case Z_OK:
                destlen = destbuf.length * 2;
                continue loop;

            case Z_STREAM_END:
                destbuf.length = zs.total_out;
                err = etc.c.zlib.inflateEnd(&zs);
                if (err != Z_OK)
                    throw new ZlibException(err);
                return destbuf;

            default:
                etc.c.zlib.inflateEnd(&zs);
                throw new ZlibException(err);
        }
    }
    assert(0, "Unreachable code");
}

@system unittest
{
    auto src =
"the quick brown fox jumps over the lazy dog\r
the quick brown fox jumps over the lazy dog\r
";
    ubyte[] dst;
    ubyte[] result;

    //arrayPrint(src);
    dst = compress(src);
    //arrayPrint(dst);
    result = cast(ubyte[]) uncompress(dst);
    //arrayPrint(result);
    assert(result == src);
}

@system unittest
{
    ubyte[] src = new ubyte[1000000];
    ubyte[] dst;
    ubyte[] result;

    src[] = 0x80;
    dst = compress(src);
    assert(dst.length*2 + 1 < src.length);
    result = cast(ubyte[]) uncompress(dst);
    assert(result == src);
}

/+
void arrayPrint(ubyte[] array)
{
    //printf("array %p,%d\n", cast(void*) array, array.length);
    for (size_t i = 0; i < array.length; i++)
    {
        printf("%02x ", array[i]);
        if (((i + 1) & 15) == 0)
            printf("\n");
    }
    printf("\n\n");
}
+/

/// the header format the compressed stream is wrapped in
enum HeaderFormat {
    deflate, /// a standard zlib header
    gzip, /// a gzip file format header
    determineFromData /// used when decompressing. Try to automatically detect the stream format by looking at the data
}

/*********************************************
 * Used when the data to be compressed is not all in one buffer.
 */

class Compress
{
    import std.conv : to;

  private:
    z_stream zs;
    int level = Z_DEFAULT_COMPRESSION;
    int inited;
    immutable bool gzip;

    void error(int err)
    {
        if (inited)
        {   deflateEnd(&zs);
            inited = 0;
        }
        throw new ZlibException(err);
    }

  public:

    /**
     * Constructor.
     *
     * Params:
     *    level = compression level. Legal values are 1 .. 9, with 1 being the least
     *            compression and 9 being the most. The default value is 6.
     *    header = sets the compression type to one of the options available
     *             in $(LREF HeaderFormat). Defaults to HeaderFormat.deflate.
     *
     * See_Also:
     *    $(LREF compress), $(LREF HeaderFormat)
     */
    this(int level, HeaderFormat header = HeaderFormat.deflate)
    in
    {
        assert(1 <= level && level <= 9, "Legal compression level are in [1, 9].");
    }
    do
    {
        this.level = level;
        this.gzip = header == HeaderFormat.gzip;
    }

    /// ditto
    this(HeaderFormat header = HeaderFormat.deflate)
    {
        this.gzip = header == HeaderFormat.gzip;
    }

    ~this()
    {   int err;

        if (inited)
        {
            inited = 0;
            deflateEnd(&zs);
        }
    }

    /**
     * Compress the data in buf and return the compressed data.
     * Params:
     *    buf = data to compress
     *
     * Returns:
     *    the compressed data. The buffers returned from successive calls to this should be concatenated together.
     *
     */
    const(void)[] compress(const(void)[] buf)
    {
        import core.memory : GC;
        import std.array : uninitializedArray;
        int err;
        ubyte[] destbuf;

        if (buf.length == 0)
            return null;

        if (!inited)
        {
            err = deflateInit2(&zs, level, Z_DEFLATED, 15 + (gzip ? 16 : 0), 8, Z_DEFAULT_STRATEGY);
            if (err)
                error(err);
            inited = 1;
        }

        destbuf = uninitializedArray!(ubyte[])(zs.avail_in + buf.length);
        zs.next_out = destbuf.ptr;
        zs.avail_out = to!uint(destbuf.length);

        if (zs.avail_in)
            buf = zs.next_in[0 .. zs.avail_in] ~ cast(ubyte[]) buf;

        zs.next_in = cast(typeof(zs.next_in)) buf.ptr;
        zs.avail_in = to!uint(buf.length);

        err = deflate(&zs, Z_NO_FLUSH);
        if (err != Z_STREAM_END && err != Z_OK)
        {
            GC.free(destbuf.ptr);
            error(err);
        }
        destbuf.length = destbuf.length - zs.avail_out;
        return destbuf;
    }

    /***
     * Compress and return any remaining data.
     * The returned data should be appended to that returned by compress().
     * Params:
     *  mode = one of the following:
     *          $(DL
                    $(DT Z_SYNC_FLUSH )
                    $(DD Syncs up flushing to the next byte boundary.
                        Used when more data is to be compressed later on.)
                    $(DT Z_FULL_FLUSH )
                    $(DD Syncs up flushing to the next byte boundary.
                        Used when more data is to be compressed later on,
                        and the decompressor needs to be restartable at this
                        point.)
                    $(DT Z_FINISH)
                    $(DD (default) Used when finished compressing the data. )
                )
     */
    void[] flush(int mode = Z_FINISH)
    in
    {
        assert(mode == Z_FINISH || mode == Z_SYNC_FLUSH || mode == Z_FULL_FLUSH,
                "Mode must be either Z_FINISH, Z_SYNC_FLUSH or Z_FULL_FLUSH.");
    }
    do
    {
        import core.memory : GC;
        ubyte[] destbuf;
        ubyte[512] tmpbuf = void;
        int err;

        if (!inited)
            return null;

        /* may be  zs.avail_out+<some constant>
         * zs.avail_out is set nonzero by deflate in previous compress()
         */
        //tmpbuf = new void[zs.avail_out];
        zs.next_out = tmpbuf.ptr;
        zs.avail_out = tmpbuf.length;

        while ( (err = deflate(&zs, mode)) != Z_STREAM_END)
        {
            if (err == Z_OK)
            {
                if (zs.avail_out != 0 && mode != Z_FINISH)
                    break;
                else if (zs.avail_out == 0)
                {
                    destbuf ~= tmpbuf;
                    zs.next_out = tmpbuf.ptr;
                    zs.avail_out = tmpbuf.length;
                    continue;
                }
                err = Z_BUF_ERROR;
            }
            GC.free(destbuf.ptr);
            error(err);
        }
        destbuf ~= tmpbuf[0 .. (tmpbuf.length - zs.avail_out)];

        if (mode == Z_FINISH)
        {
            err = deflateEnd(&zs);
            inited = 0;
            if (err)
                error(err);
        }
        return destbuf;
    }
}

/******
 * Used when the data to be decompressed is not all in one buffer.
 */

class UnCompress
{
    import std.conv : to;

  private:
    z_stream zs;
    int inited;
    int done;
    bool inputEnded;
    size_t destbufsize;

    HeaderFormat format;

    void error(int err)
    {
        if (inited)
        {   inflateEnd(&zs);
            inited = 0;
        }
        throw new ZlibException(err);
    }

  public:

    /**
     * Construct. destbufsize is the same as for D.zlib.uncompress().
     */
    this(uint destbufsize)
    {
        this.destbufsize = destbufsize;
    }

    /** ditto */
    this(HeaderFormat format = HeaderFormat.determineFromData)
    {
        this.format = format;
    }

    ~this()
    {   int err;

        if (inited)
        {
            inited = 0;
            inflateEnd(&zs);
        }
        done = 1;
    }

    /**
     * Decompress the data in buf and return the decompressed data.
     * The buffers returned from successive calls to this should be concatenated
     * together.
     */
    const(void)[] uncompress(const(void)[] buf)
    in
    {
        assert(!done, "Buffer has been flushed.");
    }
    do
    {
        if (inputEnded || !buf.length)
            return null;

        import core.memory : GC;
        import std.array : uninitializedArray;
        int err;

        if (!inited)
        {
        int windowBits = 15;
        if (format == HeaderFormat.gzip)
            windowBits += 16;
            else if (format == HeaderFormat.determineFromData)
            windowBits += 32;

            err = inflateInit2(&zs, windowBits);
            if (err)
                error(err);
            inited = 1;
        }

        if (!destbufsize)
            destbufsize = to!uint(buf.length) * 2;
        auto destbuf = uninitializedArray!(ubyte[])(destbufsize);
        size_t destFill;

        zs.next_in = cast(ubyte*) buf.ptr;
        zs.avail_in = to!uint(buf.length);

        while (true)
        {
            auto oldAvailIn = zs.avail_in;

            zs.next_out = destbuf[destFill .. $].ptr;
            zs.avail_out = to!uint(destbuf.length - destFill);

            err = inflate(&zs, Z_NO_FLUSH);
            if (err == Z_STREAM_END)
            {
                inputEnded = true;
                break;
            }
            else if (err != Z_OK)
            {
                GC.free(destbuf.ptr);
                error(err);
            }
            else if (!zs.avail_in)
                break;

            /*
                According to the zlib manual inflate() stops when either there's
                no more data to uncompress or the output buffer is full
                So at this point, the output buffer is too full
            */

            destFill = destbuf.length;

            if (destbuf.capacity)
            {
                if (destbuf.length < destbuf.capacity)
                    destbuf.length = destbuf.capacity;
                else
                {
                    auto newLength = GC.extend(destbuf.ptr, destbufsize, destbufsize);

                    if (newLength && destbuf.length < destbuf.capacity)
                        destbuf.length = destbuf.capacity;
                    else
                        destbuf.length += destbufsize;
                }
            }
            else
                destbuf.length += destbufsize;
        }

        destbuf.length = destbuf.length - zs.avail_out;
        return destbuf;
    }

    // Test for https://issues.dlang.org/show_bug.cgi?id=3191 and
    // https://issues.dlang.org/show_bug.cgi?id=9505
    @system unittest
    {
        import std.algorithm.comparison;
        import std.array;
        import std.file;
        import std.zlib;

        // Data that can be easily compressed
        ubyte[1024] originalData;

        // This should yield a compression ratio of at least 1/2
        auto compressedData = compress(originalData, 9);
        assert(compressedData.length < originalData.length / 2,
                "The compression ratio is too low to accurately test this situation");

        auto chunkSize = compressedData.length / 4;
        assert(chunkSize < compressedData.length,
                "The length of the compressed data is too small to accurately test this situation");

        auto decompressor = new UnCompress();
        ubyte[originalData.length] uncompressedData;
        ubyte[] reusedBuf;
        int progress;

        reusedBuf.length = chunkSize;

        for (int i = 0; i < compressedData.length; i += chunkSize)
        {
            auto len = min(chunkSize, compressedData.length - i);
            // simulate reading from a stream in small chunks
            reusedBuf[0 .. len] = compressedData[i .. i + len];

            // decompress using same input buffer
            auto chunk = decompressor.uncompress(reusedBuf);
            assert(progress + chunk.length <= originalData.length,
                    "The uncompressed result is bigger than the original data");

            uncompressedData[progress .. progress + chunk.length] = cast(const ubyte[]) chunk[];
            progress += chunk.length;
        }

        auto chunk = decompressor.flush();
        assert(progress + chunk.length <= originalData.length,
                "The uncompressed result is bigger than the original data");

        uncompressedData[progress .. progress + chunk.length] = cast(const ubyte[]) chunk[];
        progress += chunk.length;

        assert(progress == originalData.length,
                "The uncompressed and the original data sizes differ");
        assert(originalData[] == uncompressedData[],
                "The uncompressed and the original data differ");
    }

    @system unittest
    {
        ubyte[1024] invalidData;
        auto decompressor = new UnCompress();

        try
        {
            auto uncompressedData = decompressor.uncompress(invalidData);
        }
        catch (ZlibException e)
        {
            assert(e.msg == "data error");
            return;
        }

        assert(false, "Corrupted data didn't result in an error");
    }

    @system unittest
    {
        ubyte[2014] originalData = void;
        auto compressedData = compress(originalData, 9);

        auto decompressor = new UnCompress();
        auto uncompressedData = decompressor.uncompress(compressedData ~ cast(ubyte[]) "whatever");

        assert(originalData.length == uncompressedData.length,
                "The uncompressed and the original data sizes differ");
        assert(originalData[] == uncompressedData[],
                "The uncompressed and the original data differ");
        assert(!decompressor.uncompress("whatever").length,
                "Compression continued after the end");
    }

    /**
     * Decompress and return any remaining data.
     * The returned data should be appended to that returned by uncompress().
     * The UnCompress object cannot be used further.
     */
    void[] flush()
    in
    {
        assert(!done, "Buffer has been flushed before.");
    }
    out
    {
        assert(done, "Flushing failed.");
    }
    do
    {
        done = 1;
        return null;
    }

    /// Returns true if all input data has been decompressed and no further data
    /// can be decompressed (inflate() returned Z_STREAM_END)
    @property bool empty() const
    {
        return inputEnded;
    }

    ///
    @system unittest
    {
        // some random data
        ubyte[1024] originalData = void;

        // append garbage data (or don't, this works in both cases)
        auto compressedData = cast(ubyte[]) compress(originalData) ~ cast(ubyte[]) "whatever";

        auto decompressor = new UnCompress();
        auto uncompressedData = decompressor.uncompress(compressedData);

        assert(uncompressedData[] == originalData[],
                "The uncompressed and the original data differ");
        assert(decompressor.empty, "The UnCompressor reports not being done");
    }
}

/* ========================== unittest ========================= */

import std.random;
import std.stdio;

@system unittest // by Dave
{
    debug(zlib) writeln("std.zlib.unittest");

    bool CompressThenUncompress (void[] src)
    {
        ubyte[] dst = std.zlib.compress(src);
        double ratio = (dst.length / cast(double) src.length);
        debug(zlib) writef("src.length: %1$d, dst: %2$d, Ratio = %3$f", src.length, dst.length, ratio);
        ubyte[] uncompressedBuf;
        uncompressedBuf = cast(ubyte[]) std.zlib.uncompress(dst);
        assert(src.length == uncompressedBuf.length);
        assert(src == uncompressedBuf);

        return true;
    }


    // smallish buffers
    for (int idx = 0; idx < 25; idx++)
    {
        char[] buf = new char[uniform(0, 100)];

        // Alternate between more & less compressible
        foreach (ref char c; buf)
            c = cast(char) (' ' + (uniform(0, idx % 2 ? 91 : 2)));

        if (CompressThenUncompress(buf))
        {
            debug(zlib) writeln("; Success.");
        }
        else
        {
            return;
        }
    }

    // larger buffers
    for (int idx = 0; idx < 25; idx++)
    {
        char[] buf = new char[uniform(0, 1000/*0000*/)];

        // Alternate between more & less compressible
        foreach (ref char c; buf)
            c = cast(char) (' ' + (uniform(0, idx % 2 ? 91 : 10)));

        if (CompressThenUncompress(buf))
        {
            debug(zlib) writefln("; Success.");
        }
        else
        {
            return;
        }
    }

    debug(zlib) writefln("PASSED std.zlib.unittest");
}


@system unittest // by Artem Rebrov
{
    Compress cmp = new Compress;
    UnCompress decmp = new UnCompress;

    const(void)[] input;
    input = "tesatdffadf";

    const(void)[] buf = cmp.compress(input);
    buf ~= cmp.flush();
    const(void)[] output = decmp.uncompress(buf);

    //writefln("input = '%s'", cast(char[]) input);
    //writefln("output = '%s'", cast(char[]) output);
    assert( output[] == input[] );
}

// https://issues.dlang.org/show_bug.cgi?id=15457
@system unittest
{
    static assert(__traits(compiles, etc.c.zlib.gzclose(null)));
}