Zack's Kernel News
Into the Gaze of History, Pantless?
Linus Torvalds is not above public shaming as a way to get developers to toe a particular line. Recently when Herbert Xu submitted a merge request for some updates to the cryptographic subsystem, Linus remarked, "So perhaps somewhat ironically, the crypto tree is now the first tree I'm merging in this merge window that doesn't have a signed tag." The irony being that the signed tags are cryptographic signatures themselves. The crypto people failed to properly use crypto!
Linus went on to say:
"I don't require signed tags for kernel.org pulls, but I really do heavily prefer them, and they aren't that hard to do.
"I'm sure there are several other non-signed pull requests waiting in the queue, but still, your pull request stands out as being the first one – out of 27 so far – that didn't have it.
"Can I prod you in the direction of making signed tags a part of your workflow? The tag can contain the details of the pull – in which case git request-pull will populate the pull request with it – or it can be just some dummy message and you write the details separately in the pull request email like you do now.
"I know you have a pgp key, because I have one in my keyring from you going all the way back to 2011. And if you have lost sight of that one and need to create a new one, that [is] still better going forward than not signing your pull requests at all."
Herbert said he'd do it.
Later, in a different thread, regarding a different area of the kernel, David Howells submitted a patch via a Git tree, and Linus replied, "You have the dubious distinction of being the second pull today that didn't use a signed tag. Of 46 pulls today, only two were untagged branches, with the rest using signed tags." To which David replied that he hadn't actually intended this to be a pull request, but just an RFC. So he got out of that one!
I think Linus would probably say that he's not above any sort of contrivance to get developers to do what they're "supposed" to do. I think in a lot of cases when Linus needs to reply to a developer, he's thinking in terms of a larger question of policy, and an algorithmic take on whether, when, and how to respond to particular situations, with Linux's ongoing success as the fundamental goal. I think he took that approach long ago when he suddenly overwhelmed the GNU project, not to mention the powerful and hostile commercial computer industry in its entirety, and the expanse of the known universe with his discovery of how to lead an open source project. He's continuing to make relevant discoveries in that area all the time. It's one of the things that fascinates me about the project, and it's one reason I'm reluctant to jump on the "Linus should be nicer" bandwagon. He's been doing something amazing for 30 years, it's fundamentally experimental in nature, and I continue to be extremely curious about it, in all its tangled detail.
« Previous 1 2
Buy this article as PDF
(incl. VAT)