http://www.thelinuxdaily.com/2010/04/the-first-linux-announc...

Here's a diff between the two: http://paste.ubuntu.com/6504013/

I shouldn't disclose how this was done until it is fixed. It seems github is unhappy with how this turned out, but I hope they fix it soon. I have already written a personal apology to Linus, and also, apologies if I have offended anyone else.

>We consider correspondence sent to security@github.com our highest priority, and work to address any issues that arise as quickly as possible.

[1]: https://help.github.com/articles/responsible-disclosure-of-s...

Of course, the question is: is there any way to prevent this in a simple way? Given anyone can push the final commit, you would need some sort of commit signing, but that sounds more pain than it's really worth.

I was browsing Linus's repositories at the time and it disappeared mid-load.

https://github.com/d-snp/linux-ng/commit/b71ef5e3b1caa61dac0...

edit: I accept PR's though, let's make a big list of likes/dislikes and slap him with it ;)

Of course he does. He does not accept github pull requests, for reasons he's explained at length in, amongst others, https://github.com/torvalds/linux/pull/17

If this is true... sounds like business as usual at github. I don't get it, it's not the first time they refused to do jack shit and proactively fix reported security issues before they were used in a high-profile demo exploit.

Then again, it doesn't impact their bottom line since nobody switches or cares when that happens aside from a few days of noise, so why would they?

Apparently not:

"I shouldn’t have exploited it before reporting, so apologies if I have offended anyone."

http://vikraman.org/posts/2013/12/1/linux-ng.html

Or maybe they stole linus' key? Anyone cloned the repository before it was deleted? If so, i think the repo could help in figuring out how it was done.

Oops, docs say you're right, i didn't think of that.

WTF?

Most of the things I dislike about 'Linux' are OS-level inconsistencies, particularly that most user-land tools implement their own config file formats rather than using an existing one.

systemd is a notable exception, as it re-used the .desktop format for .service.

Good thing someone thought to fix it: https://github.com/torvalds/linux-ng/pull/4

Even though it seems like Torvalds is starting a new project, he might as well be just teasing and later pull in the Linux 3.13-14 tree.

Heh.

Are we getting trolled?

Then again, this just moves it from clever joke to irony. I'll take it.