Created
June 10, 2023 16:28
-
-
Save branleur/4dbed9de185da81f6b8eddbef3fb064e to your computer and use it in GitHub Desktop.
Dink Smallwood d2c decompressor
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
"""ddecompress - attempts to decompress .d script files""" | |
import click | |
import struct | |
import os | |
import glob | |
@click.command() | |
@click.argument("path") | |
def checkpath(path): | |
"""Checks to see if you've provided a dir or a script path""" | |
if os.path.isdir(path): | |
os.chdir(path) | |
#If you've provided a script collection it will attempt to decompress all of them | |
scripts = glob.glob("*.d") | |
for i in scripts: | |
if os.path.getsize(i) > 0: | |
decompress(i) | |
else: | |
if os.path.getsize(path) > 0: | |
decompress(path) | |
else: | |
click.echo("That file is empty") | |
def decompress(script): | |
with open(script, "rb") as f: | |
#Get the amount of pairs | |
paircount = ord(f.read(1)) - 128 | |
print("Paircount is %s" % paircount) | |
#Get the pair table (indicating what thing greater than 128 are going to expand into) | |
pairs = [struct.unpack("BB", f.read(2)) for i in range(paircount)] | |
#The byte number in the text body that corresponds to one of the pairs | |
indexes = [i + 128 for i in range(paircount)] | |
table = dict(zip(indexes, pairs)) | |
#Process it | |
outfile = scanner(f.read(), table) | |
with open(os.path.splitext(script)[0] + ".c", "w") as o: | |
click.secho("Writing outfile...", fg="yellow") | |
o.write(outfile) | |
def scanner(text, pairtable): | |
"""Scans through the text and replaces bytes with appropriate pairs""" | |
filetext = list(text) | |
#Make it do another pass until it can't find anything | |
anotherpass = 1 | |
#Store pass count for interest | |
passes = 0 | |
while anotherpass == 1: | |
anotherpass = 0 | |
passes += 1 | |
for count, data in enumerate(filetext): | |
if data > 127: | |
filetext[count] = pairtable[data][0] | |
filetext.insert(count + 1, pairtable[data][1]) | |
anotherpass = 1 | |
fileout = [chr(i) for i in filetext] | |
print("Did %s passes" % passes) | |
return ''.join(fileout) | |
if __name__ == '__main__': | |
checkpath() |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment