When I use a bzip3 compression, I would expect when block size is higher, then compressed file will be smaller, but opposite is true (which is wrong).
It is maybe not a bug report, I'm just try to think what happen actually.
I'm trying to compress Windows 7 64bit c:/windows folder (its size is about 11GB)
Is it because dictionary is stored in each independently compressed block and larger dictionary does not bring a better compression ratio, it just holds more and more place?
I can add more information if needed, I do not have enough free space to compress it and use just input tar file, but I guess it is not related to my changes.
lrzip-next Version
my version https://github.com/lukypko/lrzip-next/tree/files1
lrzip-next command line
space
What happened?
When I use a bzip3 compression, I would expect when block size is higher, then compressed file will be smaller, but opposite is true (which is wrong).
It is maybe not a bug report, I'm just try to think what happen actually. I'm trying to compress Windows 7 64bit c:/windows folder (its size is about 11GB)
Is it because dictionary is stored in each independently compressed block and larger dictionary does not bring a better compression ratio, it just holds more and more place?
lrzip-next-dev - is "my version which can process multiple files" from https://github.com/lukypko/lrzip-next/tree/files1
I can add more information if needed, I do not have enough free space to compress it and use just input tar file, but I guess it is not related to my changes.
LZMA (just for a reference)
What was expected behavior?
better compression with larger dictionary
Steps to reproduce
lazy
Relevant log output
No response
Please provide system details
x
Additional Context
No response