[bug#69780] [PATCH 4/4] DRAFT news: Add entry for ‘guix git authenticate’ changes.

[Skyler Ferris via Guix-patches via <guix-patches@gnu.org>]

[-- Attachment #1: Type: text/plain, Size: 1469 bytes --]

On 3/20/24 18:43, Skyler Ferris wrote:

> That gave me pause too, but I tested it with a pull that caused a fast-forward (no separate merge commit) and it still ran the hook. I looked for a `post-fetch` hook but couldn't find one, I agree that would be ideal.

It turns out that a post-fetch hook has been discussed on the git mailing list. Notably, someone asked about it for a very similar use-case starting at (1). Much of the discussion is about the way this person implemented the hook, because they wanted to prevent refs from being updated instead of simply alerting the user of the potentially bad commits. There is also a categorical objection to the idea of implementing a post-fetch hook (2), in part because modifying refs could lead to problems for the environment (such as clients re-fetching, then rejecting, the same commits repeatedly leading to excess bandwidth use) (3). In spite of this, there are WIP patches at the end of the thread implementing a tweak-fetch hook at the end of the thread (4), but I don't think they were ever completed... haven't looked at those too closely yet but maybe there's something somewhere else about them.

(1) https://lore.kernel.org/git/5898be69-4211-d441-494d-93477179cf0e@gaspard.io/
(2) https://lore.kernel.org/git/25bd770c-6a48-5b5d-04cc-6d02784ea3e7@gaspard.io/
(3) https://lore.kernel.org/git/87lgg1bwmi.fsf@evledraar.gmail.com/
(4) https://lore.kernel.org/git/30753d19-d77d-1a1a-ba42-afcd6fbb4223@gaspard.io/

[-- Attachment #2: Type: text/html, Size: 2480 bytes --]