Repairing Corrupted ZIP Files by Brute Force Scanning in C

+1 for questioning or ignoring commonly accepted assumptions.

Such as that compressed archives are hard if possible at all to repair. Or, say, that the optimal encoding is a very hard problem... something that David Huffman simply didn't know when he came up with his coding algorithm.

Very cool article. If you are going to have a bunch of packed structs, though, it's often easier to wrap all the structs in #pragma pack(0)/pack() rather than putting the attribute on each member. This is an MSism, but it works in gcc.

When I use my hacking skills for something like this, I sometimes fear that I am wasting time and a normal person would have solved this problem faster and easier because he isn't focused on programming all the time.

Nice to see that the fun way is the right way in this case.

That's a real hacker's work! :)

Though I recommend to read the nice post, here is the direct link to the code: http://git.zx2c4.com/BruteZip/tree/read.c

I'mn afraid your blog post is almost completely unreadable in Chrome and Opera on Windows 7.

It would be a great service to a whole lot of people for this method to be integrated into commonly used zip utilities.

I'm surprised that zip -FF wouldn't work. Sounds like it needs your patch, although it looks from the InfoZip source that the fix functionality is spread throughout the zip parsing code rather than isolated into a small routine like this one.

What would be interesting is a utility that could fix archives uploaded via FTP with ASCII mode accidentally turned on. This page claims to have a method: http://bukys.com/services/recovery/examples/

Just out of curiosity, how long did it take to process the huge zip file?

How is this different from what PKZIPFIX.EXE does?

GUI plz