--- /home/cpdev/src/classpath/java/util/zip/GZIPInputStream.java 2005-07-02 21:03:49.000000000 +0000 +++ java/util/zip/GZIPInputStream.java 2005-06-30 05:34:52.000000000 +0000 @@ -35,12 +35,10 @@ obligated to do so. If you do not wish to do so, delete this exception statement from your version. */ - package java.util.zip; -import java.io.EOFException; -import java.io.IOException; import java.io.InputStream; +import java.io.IOException; /** * This filter stream is used to decompress a "GZIP" format stream. @@ -58,32 +56,34 @@ */ public static final int GZIP_MAGIC = 0x8b1f; - /** - * The mask for bit 0 of the flag byte. - */ - static final int FTEXT = 0x1; + static final int Z_DEFLATED = 8; /** * The mask for bit 1 of the flag byte. */ - static final int FHCRC = 0x2; + static final int HEAD_CRC = 0x02; /** * The mask for bit 2 of the flag byte. */ - static final int FEXTRA = 0x4; + static final int EXTRA_FIELD = 0x04; /** * The mask for bit 3 of the flag byte. */ - static final int FNAME = 0x8; + static final int ORIG_NAME = 0x08; /** * The mask for bit 4 of the flag byte. */ - static final int FCOMMENT = 0x10; + static final int COMMENT = 0x10; /** + * The mask for all reserved bits of the flag byte. + */ + static final int RESERVED = 0xe0; + + /** * The CRC-32 checksum value for uncompressed data. */ protected CRC32 crc; @@ -94,11 +94,6 @@ protected boolean eos; /** - * Indicates whether or not the GZIP header has been read in. - */ - private boolean readGZIPHeader; - - /** * Creates a GZIPInputStream with the default buffer size. * * @param in The stream to read compressed data from @@ -127,8 +122,60 @@ throws IOException { super(in, new Inflater(true), size); + + // NOTE: header reading code taken from zlib's gzio.c. + + // Read the magic number. + int magic = eof_read() | (eof_read() << 8); + if (magic != GZIP_MAGIC) + throw new ZipException("gzip header corrupted"); + + int method = eof_read(); + int flags = eof_read(); + // Test from zlib. + if (method != Z_DEFLATED || (flags & RESERVED) != 0) + throw new ZipException("gzip header corrupted"); + + // Discard time, xflags, OS code. + for (int i = 0; i < 6; ++i) + eof_read(); + + // Skip the extra field. + if ((flags & EXTRA_FIELD) != 0) + { + int len = eof_read() | (eof_read() << 8); + while (len-- != 0) + eof_read(); + } + + if ((flags & ORIG_NAME) != 0) + { + while (true) + { + int c = eof_read(); + if (c == 0) + break; + } + } + + if ((flags & COMMENT) != 0) + { + while (true) + { + int c = eof_read(); + if (c == 0) + break; + } + } + + if ((flags & HEAD_CRC) != 0) + { + // FIXME: consider checking CRC of the header. + eof_read(); + eof_read(); + } + crc = new CRC32(); - readHeader(); } /** @@ -143,6 +190,14 @@ super.close(); } + private final int eof_read() throws IOException + { + int r = in.read(); + if (r == -1) + throw new ZipException("gzip header corrupted"); + return r & 0xff; + } + /** * Reads in GZIP-compressed data and stores it in uncompressed form * into an array of bytes. The method will block until either @@ -157,199 +212,41 @@ */ public int read(byte[] buf, int offset, int len) throws IOException { - // We first have to slurp in the GZIP header, then we feed all the - // rest of the data to the superclass. - // - // As we do that we continually update the CRC32. Once the data is - // finished, we check the CRC32. - // - // This means we don't need our own buffer, as everything is done - // in the superclass. - if (!readGZIPHeader) - readHeader(); - if (eos) return -1; - - // System.err.println("GZIPIS.read(byte[], off, len ... " + offset + " and len " + len); - - /* We don't have to read the header, - * so we just grab data from the superclass. - */ - int numRead = super.read(buf, offset, len); - if (numRead > 0) - crc.update(buf, offset, numRead); - - if (inf.finished()) - readFooter(); - return numRead; - } - - - /** - * Reads in the GZIP header. - */ - private void readHeader() throws IOException - { - /* 1. Check the two magic bytes */ - CRC32 headCRC = new CRC32(); - int magic = in.read(); - if (magic < 0) - { - eos = true; - return; - } - int magic2 = in.read(); - if ((magic + (magic2 << 8)) != GZIP_MAGIC) - throw new IOException("Error in GZIP header, bad magic code"); - headCRC.update(magic); - headCRC.update(magic2); - - /* 2. Check the compression type (must be 8) */ - int CM = in.read(); - if (CM != 8) - throw new IOException("Error in GZIP header, data not in deflate format"); - headCRC.update(CM); - - /* 3. Check the flags */ - int flags = in.read(); - if (flags < 0) - throw new EOFException("Early EOF in GZIP header"); - headCRC.update(flags); - - /* This flag byte is divided into individual bits as follows: - - bit 0 FTEXT - bit 1 FHCRC - bit 2 FEXTRA - bit 3 FNAME - bit 4 FCOMMENT - bit 5 reserved - bit 6 reserved - bit 7 reserved - */ - - /* 3.1 Check the reserved bits are zero */ - if ((flags & 0xd0) != 0) - throw new IOException("Reserved flag bits in GZIP header != 0"); - - /* 4.-6. Skip the modification time, extra flags, and OS type */ - for (int i=0; i< 6; i++) - { - int readByte = in.read(); - if (readByte < 0) - throw new EOFException("Early EOF in GZIP header"); - headCRC.update(readByte); - } - - /* 7. Read extra field */ - if ((flags & FEXTRA) != 0) - { - /* Skip subfield id */ - for (int i=0; i< 2; i++) + int r = super.read(buf, offset, len); + if (r == -1) { - int readByte = in.read(); - if (readByte < 0) - throw new EOFException("Early EOF in GZIP header"); - headCRC.update(readByte); - } - if (in.read() < 0 || in.read() < 0) - throw new EOFException("Early EOF in GZIP header"); - - int len1, len2, extraLen; - len1 = in.read(); - len2 = in.read(); - if ((len1 < 0) || (len2 < 0)) - throw new EOFException("Early EOF in GZIP header"); - headCRC.update(len1); - headCRC.update(len2); + eos = true; - extraLen = (len1 << 8) | len2; - for (int i = 0; i < extraLen;i++) - { - int readByte = in.read(); - if (readByte < 0) - throw new EOFException("Early EOF in GZIP header"); - headCRC.update(readByte); + byte[] tmp = new byte[8]; + // First copy remaining bytes from inflater input buffer. + int avail = inf.getRemaining(); + System.arraycopy(this.buf, this.len - avail, tmp, 0, avail); + + // Now read remaining bytes from wrapped input stream. + for (int i = avail; i < 8; ++i) + { + tmp[i] = (byte) eof_read(); + } + + // Be careful to avoid sign extension here; CRC32.getValue() + // returns a long. + long header_crc = read4(tmp, 0) & 0xffffffffL; + if (crc.getValue() != header_crc) + throw new ZipException("corrupted gzip file - crc mismatch"); + int isize = read4(tmp, 4); + if (inf.getTotalOut() != isize) + throw new ZipException("corrupted gzip file - size mismatch"); + return -1; } - } - - /* 8. Read file name */ - if ((flags & FNAME) != 0) - { - int readByte; - while ( (readByte = in.read()) > 0) - headCRC.update(readByte); - if (readByte < 0) - throw new EOFException("Early EOF in GZIP file name"); - headCRC.update(readByte); - } - - /* 9. Read comment */ - if ((flags & FCOMMENT) != 0) - { - int readByte; - while ( (readByte = in.read()) > 0) - headCRC.update(readByte); - - if (readByte < 0) - throw new EOFException("Early EOF in GZIP comment"); - headCRC.update(readByte); - } - - /* 10. Read header CRC */ - if ((flags & FHCRC) != 0) - { - int tempByte; - int crcval = in.read(); - if (crcval < 0) - throw new EOFException("Early EOF in GZIP header"); - - tempByte = in.read(); - if (tempByte < 0) - throw new EOFException("Early EOF in GZIP header"); - - crcval = (crcval << 8) | tempByte; - if (crcval != ((int) headCRC.getValue() & 0xffff)) - throw new IOException("Header CRC value mismatch"); - } - - readGZIPHeader = true; - //System.err.println("Read GZIP header"); + crc.update(buf, offset, r); + return r; } - private void readFooter() throws IOException + private final int read4(byte[] buf, int offset) throws IOException { - byte[] footer = new byte[8]; - int avail = inf.getRemaining(); - if (avail > 8) - avail = 8; - System.arraycopy(buf, len - inf.getRemaining(), footer, 0, avail); - int needed = 8 - avail; - while (needed > 0) - { - int count = in.read(footer, 8-needed, needed); - if (count <= 0) - throw new EOFException("Early EOF in GZIP footer"); - needed -= count; //Jewel Jan 16 - } - - int crcval = (footer[0] & 0xff) | ((footer[1] & 0xff) << 8) - | ((footer[2] & 0xff) << 16) | (footer[3] << 24); - if (crcval != (int) crc.getValue()) - throw new IOException("GZIP crc sum mismatch, theirs \"" - + Integer.toHexString(crcval) - + "\" and ours \"" - + Integer.toHexString( (int) crc.getValue())); - - int total = (footer[4] & 0xff) | ((footer[5] & 0xff) << 8) - | ((footer[6] & 0xff) << 16) | (footer[7] << 24); - if (total != inf.getTotalOut()) - throw new IOException("Number of bytes mismatch"); - - /* FIXME" XXX Should we support multiple members. - * Difficult, since there may be some bytes still in buf - */ - eos = true; + return (((buf[offset + 3] & 0xFF) << 24) + ((buf[offset + 2] & 0xFF) << 16) + + ((buf[offset + 1] & 0xFF) << 8) + (buf[offset] & 0xFF)); } }