jslzjb-k
This is a fork of the jslzjb project, https://code.google.com/p/jslzjb/
It has been modified to work on Uint8Array values exclusively and avoids all
allocations during compression and decompression. These changes have been made
out of the need to compress large buffers in a reasonable amount of time.
It performs compression of a 100 MB buffer in the order of a second.
Test it yourself:
% ./test-lzjb.js test-lzjb.js
compress: 0.822ms
decompress: 0.281ms
compressed test-lzjb.js from=905 to=659
% ./test-lzjb.js /bin/node
compress: 106.155ms
decompress: 55.874ms
compressed /bin/node from=14059k to=9348k
Two methods are exported:
/**
* Compress byte array using fast and efficient algorithm.
*
* @param {Uint8Array} sstart The buffer to compress
* @param {Uint8Array} dstart The buffer to write into
* @return {number} compressed length (number of bytes written to the
* output buffer). May be bigger than the size of the
* output buffer, in which case some bytes are lost
*/
jslzjb.compress = function(sstart, dstart)
/**
* Decompress byte array using fast and efficient algorithm.
*
* @param {Uint8Array} sstart The buffer to decompress
* @param {number} slen compressed length
* @param {Uint8Array} dstart The buffer to write into
* @return {number} decompressed length
*/
jslzjb.decompress = function(sstart, slen, dstart)TODO
- Benchmarks
- Make use of asm.js
- Make use of simd