mirror of
https://github.com/corda/corda.git
synced 2025-01-07 13:38:47 +00:00
d4da92d300
Previously, Deflater.deflate would pass Z_SYNC_FLUSH to zlib unconditionally, which caused the output to be enormous when setInput was called repeatedly with very small input buffers. In order to allow zlib to buffer output and thereby maximize compression, we must use Z_NO_FLUSH until Deflater.finish is called, at which point we switch to Z_FINISH. We also modify DeflaterOutputStream.close to call Deflater.finish and write any remaining output to the wrapped stream. |
||
---|---|---|
.. | ||
avian | ||
java | ||
sun | ||
java-io.cpp | ||
java-lang.cpp | ||
java-net.cpp | ||
java-nio.cpp | ||
java-util-zip.cpp | ||
java-util.cpp | ||
jni-util.h |