comand uses IPC through shared memory and Huffman coding to compress and decompress data. No third-party libraries are used, only pure C.
This project is composed of two programs: encoder and decoder. encoder reads the data, compress it and shares it using a shared memory segment. decoder will decompress the compressed data, sending it back to the first program - which will print the received stats.
Since it was an one-week challenge, comand is a relatively small program:
.
├── Makefile
├── include
│ └── comand.h
├── lib
│ ├── comand_ipc
│ │ ├── c_ipc.c
│ │ ├── c_ipc.h
│ │ └── Makefile
│ └── huffman
│ ├── huffman_ds.c
│ ├── huffman_ds.h
│ └── Makefile
└── src
├── decoder
│ ├── decoder.c
│ ├── decoder.h
│ ├── decompress.c
│ └── Makefile
└── encoder
├── compress.c
├── encoder.c
├── encoder.h
├── input.c
└── MakefileAfter cloning this repository, run:
makeYou can now run encoder:
./src/encoder/encoder "text to be " "compressed " "!!!!!"
# Alternatively, run it with '--file' option, to compress
# the contents of a given file
./src/encoder/encoder --file file_to_have_its_contents_compressedAfter compressing the data, it will send it to decoder and start to wait for feedback, so you can run decoder:
./src/decoder/decoderencoder will print received stats such as the decompressed data and its total bytes, total of compressd bytes and time used to decompress.
comand has some limitations, of course. There are two particular problems regarding the input size.
First of all, with a single letter input decoder is not capable of decompressing it to its original form. I've chosen not to treat it.
./encoder "a" # ./decoder will not decompress it properlyAlso, my implementation uses a fixed array size to store the "Huffman tree", due to the difficulty of using pointers inside shared memory (again, this is an one-week challenge). So it means that some enormous input may not work properly.
comand is licensed under the terms of the GPL License.