I have a friend who works from home. He produces files that he needs to upload to the company's server. He uses a Cisco AnyConnect VPN to connect, and then drags/drops the files to the server. To reduce transfer time, he typically compresses them with ZIP. People complain that they cannot open the files because they are corrupt.
I had him create a test file by ZIPping some random files. The test file was about 2MB. I had him transfer the file to the server and then transfer it back. The file transferred back did not match the transferred file. Looking at both files, they have the exact same number of bytes and match for the first 700KB or so. After that, chunks of the file differ. The data hasn't been replaced by zeroes or subtly corrupted -- in the chunks that differ, the data in the server-side file is just completely different.
His machine is a Mac. The server is Linux. He drags/drops the files to transfer them. They're just ZIP files, so I don't think it's a resource fork issue.
This used to work, but it's not exactly clear when it started not working, so it's hard to track down what might have changed. Any ideas for possible avenues to pursue?