aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2017-04-19Update home page URL to HTTPS.Lasse Collin
2016-11-27Bump the version number to 1.6.Lasse Collin
2016-11-27Add NEWS for 1.6.Lasse Collin
2016-10-26Update THANKS.Lasse Collin
2016-10-26Support known uncompressed size in LZMAOutputStream.Lasse Collin
This allows storing the uncompressed size to the .lzma header and omitting the end of stream marker. Thanks to Matthias Stevens for the initial patch and testing.
2016-09-10Improve an error message about corrupt XZ Index.Lasse Collin
2016-09-10Clean up casts a little in the range encoder classes.Lasse Collin
OutputStream.write(int) has an int argument so it's useless to cast to byte in shiftLow().
2016-09-10Add LZMAEncDemo.Lasse Collin
2016-09-10Add LZMAOutputStream for .lzma format and raw LZMA streams.Lasse Collin
2016-09-10Support unsigned 32-bit values in LZMAEncoder.getDistSlot.Lasse Collin
This will help when writing the LZMA1 end of payload marker.
2016-09-10Add RangeEncoderToStream.Lasse Collin
2016-09-10Allow RangeEncoder.writeByte() to throw an IOException.Lasse Collin
2016-09-10Split RangeEncoderToBuffer from RangeEncoder.Lasse Collin
2015-12-04Update THANKS.Lasse Collin
2015-12-04Fix LZMA2Options.getInputStream to work with a preset dictionary.Lasse Collin
Thanks to Igor Sakovich.
2014-08-14Make it possible to disable verification of integrity checks.Lasse Collin
This feature should almost never be used but there are situations where it is useful.
2014-08-14Refactor the constructors of SingleXZInputStream.Lasse Collin
2014-04-24Make a few fields final.Lasse Collin
Thanks to Carl Hasselskog.
2014-03-08Bump the version number to 1.5.Lasse Collin
2014-03-08Add NEWS for 1.5.Lasse Collin
2014-03-08Rename tmpbuf to filterBuf.Lasse Collin
In the commit a9bf23ff908ac7f8f4df25fd621e04e187425a7d the field tempBuf was added to many classes. Three classes already had tmpbuf and TMPBUF_SIZE fields which sound too similar, so it's good to rename them.
2014-03-06Remove an unneeded field from BlockInputStream.Lasse Collin
2014-03-06Fix coding style in CRC64.java.Lasse Collin
2014-03-06Omit a useless use of new.Lasse Collin
2014-03-03Use fields for temp buffers in read() and write(int).Lasse Collin
This avoids reallocating memory for each call. This reduces pressure on the garbage collector and in practice lowers the memory used by the java process significantly when read() and write(int) functions are used a lot. On the other hand, this is stupid since most of the time the read() and write(int) aren't (hopefully) used since one gets better performance with functions that process arrays of bytes. Thus, for most uses cases this commit means that we are allocating a few unused one-byte buffers that weren't previously allocated at all. Similar thing was done earlier in the LZMA encoder in the commit ec224582e44776a53874346dfe703dbbfcc6bd15.
2014-02-28Fix a wrong assertion.Lasse Collin
Thanks to Stefan Bodewig.
2013-09-22Bump the version number to 1.4.Lasse Collin
2013-09-22Add NEWS for 1.4.Lasse Collin
2013-05-30Add LZMADecDemo demo program.Lasse Collin
2013-05-30Add LZMAInputStream to support .lzma files and raw LZMA1 streams.Lasse Collin
2013-05-30Split RangeDecoder into RangeDecoderFromBuffer and FromStream.Lasse Collin
To support LZMA1, support for reading from an InputStream is needed. LZMA1-specific features are in FromStream while LZMA2-specific features are in FromBuffer. Move the rc.isInBufferOK() from LZMADecoder to LZMA2InputStream since that is a LZMA2-specific thing.
2013-05-30Comment improvements to LZMA2InputStream.Lasse Collin
2013-05-12Bump the version number to 1.3.Lasse Collin
2013-05-12Update the external documentation URL to point to Java 7 docs.Lasse Collin
2013-05-12Make the reference to GZIPInputStream a link.Lasse Collin
2013-05-12Add NEWS for 1.3.Lasse Collin
2013-05-11Fix data corruption in LZMA2 encoder with flushing or preset dictionary.Lasse Collin
Process pending match finder data when flushing or finishing is requested. Previously it was done only when new data was added and only if enough data was added to make it possible for the encoder to encode at least one byte. But if flushing or finishing was requested after adding a few (but not enough) bytes, the pending data wasn't processed and the match finder got out of sync (and could make an assertion fail if assertions were enabled). Bug report: http://sourceforge.net/projects/lzmautils/forums/forum/708858/topic/7796400
2013-05-09Fix a comment.Lasse Collin
2013-04-17Add seekToBlock to SeekableXZInputStream.Lasse Collin
2013-04-14Add getBlockNumber to SeekableXZInputStream.Lasse Collin
2013-04-14Make Block positions and sizes public in SeekableXZInputStream.Lasse Collin
This adds the functions getBlockPos, getBlockSize, getBlockCompPos, getBlockCompSize, and getBlockCheckType.
2013-04-14Comment improvements to SeekableXZInputStream.Lasse Collin
2013-04-14Show the Stream and Block counts in XZSeekDecDemo.Lasse Collin
2013-04-14Add getStreamCount and getBlockCount to SeekableXZInputStream.Lasse Collin
2013-04-14Restructure SeekableXZInputStream and IndexDecoder.Lasse Collin
Now the BlockInfo about the current Block is held in SeekableXZInputStream instead of having a BlockInfo in every IndexDecoder in a multi-Stream .xz file. This will be useful in the next commits that make the Block boundaries visible outside the SeekableXZInputStream.
2013-04-14Comment improvements.Lasse Collin
2013-04-14Silence a compiler warning.Lasse Collin
2013-01-29Bump the version number to 1.2.Lasse Collin
2013-01-29Add NEWS for 1.2.Lasse Collin
2013-01-29Update THANKS.Lasse Collin