Freearc with Reflate test

[Razor12911]

Been obsessed with reflate that I ended up with two versions. Anyways with the help from Shegorat with the source, this is where I ended up.

Normal reflate (reflate.exe)
Raw2Hif c6
Method -mreflate+lzma
Streams detected 60375
Input 858MB
Output 307MB
Decompression speed 870KB/s using reflate.exe

Advantages with normal reflate
Compression speed is faster than modded reflate.
More information is produced about compression.
Suitable for CMD work.

Disadvantages with normal reflate
Output is bigger than modded reflate although the method and input is the same.
Progress cannot be determined without cmd window.
Needs more space when decompressing.
Decompression speed is slower than modded reflate.

Modded reflate (cls-reflate.dll)
Raw2Hif c6
Method -mreflate+lzma
Streams detected 60375
Input 858MB (Same file)
Output 291MB
Decompression speed 1,619KB/s using cls-reflate.dll

Advantages with modded reflate
Shows progress while unpacking.
Output is smaller than normal reflate
Needs less space when decompressing.
It's faster than normal reflate on some occasions.

Disadvantages with modded reflate
Needs more space when compressing.
Compression speed is slightly slower than normal reflate
Quite buggy, more tests have to be run.
Cannot be terminated once started decompression has started. (Really have fix this)
It can make your computer freeze.
Still has to be optimised.
It's not ready for public use.

What keeps me wondering is the size difference between normal and modded reflate, same input but different output although the same method is used, the file is perfectly restored because CRC check is the same throughout. Although I think I know why.


Tested cls-reflate.dll and it works with Windows Phone Installer.