cbzip2outputstream.java

java ant的源码！非常值得看的源码

        /* 20 is just a paranoia constant */
        this.allowableBlockSize
            = (this.blockSize100k * BZip2Constants.baseBlockSize) - 20;
    }

    private void endBlock() throws IOException {
        this.blockCRC = this.crc.getFinalCRC();
        this.combinedCRC = (this.combinedCRC << 1) | (this.combinedCRC >>> 31);
        this.combinedCRC ^= this.blockCRC;

        // empty block at end of file
        if (this.last == -1) {
            return;
        }

        /* sort the block and establish posn of original string */
        blockSort();

        /*
          A 6-byte block header, the value chosen arbitrarily
          as 0x314159265359 :-).  A 32 bit value does not really
          give a strong enough guarantee that the value will not
          appear by chance in the compressed datastream.  Worst-case
          probability of this event, for a 900k block, is about
          2.0e-3 for 32 bits, 1.0e-5 for 40 bits and 4.0e-8 for 48 bits.
          For a compressed file of size 100Gb -- about 100000 blocks --
          only a 48-bit marker will do.  NB: normal compression/
          decompression do *not* rely on these statistical properties.
          They are only important when trying to recover blocks from
          damaged files.
        */
        bsPutUByte(0x31);
        bsPutUByte(0x41);
        bsPutUByte(0x59);
        bsPutUByte(0x26);
        bsPutUByte(0x53);
        bsPutUByte(0x59);

        /* Now the block's CRC, so it is in a known place. */
        bsPutInt(this.blockCRC);

        /* Now a single bit indicating randomisation. */
        if (this.blockRandomised) {
            bsW(1, 1);
        } else {
            bsW(1, 0);
        }

        /* Finally, block's contents proper. */
        moveToFrontCodeAndSend();
    }

    private void endCompression() throws IOException {
        /*
          Now another magic 48-bit number, 0x177245385090, to
          indicate the end of the last block.  (sqrt(pi), if
          you want to know.  I did want to use e, but it contains
          too much repetition -- 27 18 28 18 28 46 -- for me
          to feel statistically comfortable.  Call me paranoid.)
        */
        bsPutUByte(0x17);
        bsPutUByte(0x72);
        bsPutUByte(0x45);
        bsPutUByte(0x38);
        bsPutUByte(0x50);
        bsPutUByte(0x90);

        bsPutInt(this.combinedCRC);
        bsFinishedWithStream();
    }

    /**
     * Returns the blocksize parameter specified at construction time.
     */
    public final int getBlockSize() {
        return this.blockSize100k;
    }

    public void write(final byte[] buf, int offs, final int len)
        throws IOException {
        if (offs < 0) {
            throw new IndexOutOfBoundsException("offs(" + offs + ") < 0.");
        }
        if (len < 0) {
            throw new IndexOutOfBoundsException("len(" + len + ") < 0.");
        }
        if (offs + len > buf.length) {
            throw new IndexOutOfBoundsException("offs(" + offs + ") + len("
                                                + len + ") > buf.length("
                                                + buf.length + ").");
        }
        if (this.out == null) {
            throw new IOException("stream closed");
        }

        for (int hi = offs + len; offs < hi;) {
            write0(buf[offs++]);
        }
    }

    private void write0(int b) throws IOException {
        if (this.currentChar != -1) {
            b &= 0xff;
            if (this.currentChar == b) {
                if (++this.runLength > 254) {
                    writeRun();
                    this.currentChar = -1;
                    this.runLength = 0;
                }
                // else nothing to do
            } else {
                writeRun();
                this.runLength = 1;
                this.currentChar = b;
            }
        } else {
            this.currentChar = b & 0xff;
            this.runLength++;
        }
    }

    private static void hbAssignCodes(final int[] code, final byte[] length,
                                      final int minLen, final int maxLen,
                                      final int alphaSize) {
        int vec = 0;
        for (int n = minLen; n <= maxLen; n++) {
            for (int i = 0; i < alphaSize; i++) {
                if ((length[i] & 0xff) == n) {
                    code[i] = vec;
                    vec++;
                }
            }
            vec <<= 1;
        }
    }

    private void bsFinishedWithStream() throws IOException {
        while (this.bsLive > 0) {
            int ch = this.bsBuff >> 24;
            this.out.write(ch); // write 8-bit
            this.bsBuff <<= 8;
            this.bsLive -= 8;
        }
    }

    private void bsW(final int n, final int v) throws IOException {
        final OutputStream outShadow = this.out;
        int bsLiveShadow    = this.bsLive;
        int bsBuffShadow    = this.bsBuff;

        while (bsLiveShadow >= 8) {
            outShadow.write(bsBuffShadow >> 24); // write 8-bit
            bsBuffShadow <<= 8;
            bsLiveShadow -= 8;
        }

        this.bsBuff = bsBuffShadow | (v << (32 - bsLiveShadow - n));
        this.bsLive = bsLiveShadow + n;
    }

    private void bsPutUByte(final int c) throws IOException {
        bsW(8, c);
    }

    private void bsPutInt(final int u) throws IOException {
        bsW(8, (u >> 24) & 0xff);
        bsW(8, (u >> 16) & 0xff);
        bsW(8, (u >>  8) & 0xff);
        bsW(8,  u        & 0xff);
    }

    private void sendMTFValues() throws IOException {
        final byte[][] len  = this.data.sendMTFValues_len;
        final int alphaSize = this.nInUse + 2;

        for (int t = N_GROUPS; --t >= 0;) {
            byte[] len_t = len[t];
            for (int v = alphaSize; --v >= 0;) {
                len_t[v] = GREATER_ICOST;
            }
        }

        /* Decide how many coding tables to use */
        // assert (this.nMTF > 0) : this.nMTF;
        final int nGroups =
            (this.nMTF <  200) ? 2
            : (this.nMTF <  600) ? 3
            : (this.nMTF < 1200) ? 4
            : (this.nMTF < 2400) ? 5
            : 6;

        /* Generate an initial set of coding tables */
        sendMTFValues0(nGroups, alphaSize);

        /*
          Iterate up to N_ITERS times to improve the tables.
        */
        final int nSelectors = sendMTFValues1(nGroups, alphaSize);

        /* Compute MTF values for the selectors. */
        sendMTFValues2(nGroups, nSelectors);

        /* Assign actual codes for the tables. */
        sendMTFValues3(nGroups, alphaSize);

        /* Transmit the mapping table. */
        sendMTFValues4();

        /* Now the selectors. */
        sendMTFValues5(nGroups, nSelectors);

        /* Now the coding tables. */
        sendMTFValues6(nGroups, alphaSize);

        /* And finally, the block data proper */
        sendMTFValues7(nSelectors);
    }

    private void sendMTFValues0(final int nGroups, final int alphaSize) {
        final byte[][] len  = this.data.sendMTFValues_len;
        final int[] mtfFreq = this.data.mtfFreq;

        int remF = this.nMTF;
        int gs = 0;

        for (int nPart = nGroups; nPart > 0; nPart--) {
            final int tFreq = remF / nPart;
            int ge = gs - 1;
            int aFreq = 0;

            for (final int a = alphaSize - 1; (aFreq < tFreq) && (ge < a);) {
                aFreq += mtfFreq[++ge];
            }

            if ((ge > gs)
                && (nPart != nGroups)
                && (nPart != 1)
                && (((nGroups - nPart) & 1) != 0)) {
                aFreq -= mtfFreq[ge--];
            }

            final byte[] len_np = len[nPart - 1];
            for (int v = alphaSize; --v >= 0;) {
                if ((v >= gs) && (v <= ge)) {
                    len_np[v] = LESSER_ICOST;
                } else {
                    len_np[v] = GREATER_ICOST;
                }
            }

            gs = ge + 1;
            remF -= aFreq;
        }
    }

    private int sendMTFValues1(final int nGroups, final int alphaSize) {
        final Data dataShadow       = this.data;
        final int[][] rfreq   = dataShadow.sendMTFValues_rfreq;
        final int[] fave      = dataShadow.sendMTFValues_fave;
        final short[] cost    = dataShadow.sendMTFValues_cost;
        final char[] sfmap    = dataShadow.sfmap;
        final byte[] selector = dataShadow.selector;
        final byte[][] len    = dataShadow.sendMTFValues_len;
        final byte[] len_0 = len[0];
        final byte[] len_1 = len[1];
        final byte[] len_2 = len[2];
        final byte[] len_3 = len[3];
        final byte[] len_4 = len[4];
        final byte[] len_5 = len[5];
        final int nMTFShadow = this.nMTF;

        int nSelectors = 0;

        for (int iter = 0; iter < N_ITERS; iter++) {
            for (int t = nGroups; --t >= 0;) {
                fave[t] = 0;
                int[] rfreqt = rfreq[t];
                for (int i = alphaSize; --i >= 0;) {
                    rfreqt[i] = 0;
                }
            }

            nSelectors = 0;

            for (int gs = 0; gs < this.nMTF;) {
                /* Set group start & end marks. */

                /*
                  Calculate the cost of this group as coded
                  by each of the coding tables.
                */

                final int ge = Math.min(gs + G_SIZE - 1, nMTFShadow - 1);

                if (nGroups == N_GROUPS) {
                    // unrolled version of the else-block

                    short cost0 = 0;
                    short cost1 = 0;
                    short cost2 = 0;
                    short cost3 = 0;
                    short cost4 = 0;
                    short cost5 = 0;

                    for (int i = gs; i <= ge; i++) {
                        final int icv = sfmap[i];
                        cost0 += len_0[icv] & 0xff;
                        cost1 += len_1[icv] & 0xff;
                        cost2 += len_2[icv] & 0xff;
                        cost3 += len_3[icv] & 0xff;
                        cost4 += len_4[icv] & 0xff;
                        cost5 += len_5[icv] & 0xff;
                    }

                    cost[0] = cost0;
                    cost[1] = cost1;
                    cost[2] = cost2;
                    cost[3] = cost3;
                    cost[4] = cost4;
                    cost[5] = cost5;

                } else {
                    for (int t = nGroups; --t >= 0;) {
                        cost[t] = 0;
                    }

                    for (int i = gs; i <= ge; i++) {
                        final int icv = sfmap[i];
                        for (int t = nGroups; --t >= 0;) {
                            cost[t] += len[t][icv] & 0xff;
                        }
                    }
                }

                /*
                  Find the coding table which is best for this group,
                  and record its identity in the selector table.
                */
                int bt = -1;
                for (int t = nGroups, bc = 999999999; --t >= 0;) {
                    final int cost_t = cost[t];
                    if (cost_t < bc) {
                        bc = cost_t;
                        bt = t;
                    }
                }

                fave[bt]++;
                selector[nSelectors] = (byte) bt;
                nSelectors++;

                /*
                  Increment the symbol frequencies for the selected table.
                */
                final int[] rfreq_bt = rfreq[bt];
                for (int i = gs; i <= ge; i++) {
                    rfreq_bt[sfmap[i]]++;
                }

                gs = ge + 1;
            }

            /*
              Recompute the tables based on the accumulated frequencies.
            */
            for (int t = 0; t < nGroups; t++) {
                hbMakeCodeLengths(len[t], rfreq[t], this.data, alphaSize, 20);
            }
        }

        return nSelectors;
    }

    private void sendMTFValues2(final int nGroups, final int nSelectors) {
        // assert (nGroups < 8) : nGroups;

        final Data dataShadow = this.data;
        byte[] pos = dataShadow.sendMTFValues2_pos;

        for (int i = nGroups; --i >= 0;) {
            pos[i] = (byte) i;
        }

        for (int i = 0; i < nSelectors; i++) {