Thread: test-me: text compression

also, a question:
this program is supposed to be able to compress anything in bytes (ie, binaries, too). but for some reason, after decompression of a binary file like an image or program, the file's much smaller, and corrupt. any idea on why this is happening?

howdy,
i tryed it on RH 7.2 kernel 2.4.7-10 using gcc 2.96.
it worked just fine, i had a file that started at 2247 before copression and after compression it was 1512 not counting the look up table. when i decompressed it it came just like the original.
now i will disect your source code and steal any knowledge from it that i can.
very very cool thanks for sharing your stuff

M.R.

the file you compressed, was it a text file or a binary file?

go ahead, steal. it's a huffman algorithm. in the future i'm going to change the table so it will compress smaller.
a little help:

class huffman_tree // a class with a vector array of nodes, and member functions to sort, search, organize into a huffman tree, etc

class huff_table // a lookup table, basically a vector array for the letters, a vector<bool> for each binary, and member functions to encode and decode a string to it

struct node // a node. it has a value (the letter), an amount (the frequency of the letter) and a pointer going up, and two going down the tree.
if you are going to mess around with it please send me any changes, so i can make this better.

howdy,
the file i compressed was a file of random words and text i created while i was talking on the phone.
i created it in emacs and saved it with .txt extension. i'm not perfectly sure but i believe emacs saves this type of file as ascii text.

M.R.

i'm not perfectly sure but i believe emacs saves this type of file as ascii text.

yeah, i think so too. as stated before, binaries should but don't work. maybe it's the number of different ascii chars.