Hacker News Reader

Comment by chatmasta | Hacker News Reader

Aww that’s disappointing. Fun project.

Is this the code? https://github.com/gudvardur/amazon_book_downloader

Yes. FWIW, as of a few minutes ago when I cloned this one, all the non-git files have the same hashes as the copy of the original I cloned when it was still up.

(to clarify, I wasn't talking about any git-specific hashes, just regular sha2/blake2b hashes of python, json, and font files. However, the two sha1 commit hashes in the git history match as well.)

This is a weird thing with how GitHub forks work. All the objects within a fork network are stored within a global namespace, so you can change the repository name in the URL and find objects that appear to belong to one repository despite being unique to a fork.

These commit sha's do not match the original:

4526863 - Initial commit

Where are you getting that hex prefix from?

Sorry, I don't comment very often and not trolling. I had GitHub open to the repo on my phone and seeing that it had been taken down grabbed a screenshot of the page https://imgur.com/a/IzUA8mP

The HN thread began on Oct 16 at 20:22 Z. If you visited that github page instantly and took that screenshot instantly, even accounting for 21 hours due to rounding, the commit in that screenshot had to be after Oct 15 23:22.

The repo as I and many other people cloned it has the first commit ("first commit", not "initial commit") at Oct 12 23:20 Z, and the "done" commit at Oct 15 19:37 Z.

A likely explanation is that pixelmelt squashed both commits at or after they put up the blog post, but didn't force-push the rewritten history to github until it hit HN and blew up.

That's plausible. In the version of the repository I saw there was a GPL license file. The new repository does not have that.

Reply on news.ycombinator.com