Closed peacherwu closed 3 years ago
I was able to implement a input stream buffer reading from decompressed file one byte at a time. Also I reduced the output buffer to 1K. comp.dict_size
is set to 1K also at compression time.
However, the valgrind massif
program is still reporting a 15K memory usage. All Makefiles are the original. Is that normal?
Decompression streaming is supported, as clearly stated in the readme. Usage example can be found in included tgunzip.c. I'm not sure about other things, missing context/details.
Similar to
#2
, I am looking for a way to stream compressed file for decompression.I understand there's discussion as to how much the input buffer would be. I wonder how this relates to
dict_size
andhash_bits
set at compression time?