Live data from Hacker News

How we decreased GitLab repo backup times from 48 hours to 41 minutes

about.gitlab.com

1–10 of 273 posts

Re: How we decreased GitLab repo backup times from 48 hours to 41 minutes

#5
post #2

See also: https://www.tumblr.com/accidentallyquadratic Quadratic complexity sits in an awkward sweet spot: Fast enough for medium-sized n to pass first QA, but doomed to fail eventually as n grows.

This particular change was not accidental. It was announced as quadratic.

Re: How we decreased GitLab repo backup times from 48 hours to 41 minutes

#7
post #3

Are there any reimplementations of git, by professional programmers using real tools? The source in question — object.c — is "banging rocks together" material.

Ignoring your inaccurate flamebait and answering the underlying question: there are several reimplementations of git, including "got" (Game of Trees) from the OpenBSD developers, jgit (reimplementation in Java), the Haskell git package, the gitoxide rust crate (and I assume several more half-finished Rust hobby projects), and probably others; that's just what I found with a cursory google.

Re: How we decreased GitLab repo backup times from 48 hours to 41 minutes

#8
post #6

"fixed it with an algorithmic change, reducing backup times exponentially" If the backup times were O(n^2), are they now O(n^2 / 2^n)? I would guess not.

This is not the precise mathematical definition of exponential, but rather the colloquial one, where it just means "a lot".
Post reply on HN