Magical rust blog

Have the unicorns taken over?

Or was control given away.

1 Like

I'd say it was a combination of both.

1 Like

What is this referring to?

1 Like

I just got a github http error 500 internal server error (which has a picture of a unicorn) while trying to load the inside blog. Refreshing it made it go away, but presumably OP had this happen to them too.

For anyone who hasn't seen the error page before:

uncorn

Yeah.
I know it is "volunteer" but would hope that someone would fix that github lock-in. I always figured it was just "temporary".

I know it is too much work, to much money, and probably impossible. But shouldn't the rust blogs, books, src tarballs be put out as something like magnet links.

I actually had no idea what you were referring to, I thought we were brainstorming ideas for a short story

5 Likes

While the rust-lang blog hosted using GitHub Pages, the hosting is of a generated static site (this is all GitHub offers). So the lock-in there is no different than using any other webhost; as long as some contributor has a copy of the git repo (a distinct advantage of git being distributed), redeploying to a different service if GitHub disappears is as straightforward as it could be.

The cratesio database offers regular data dumps and all toolchain releases are published as tarballs. Disaster recovery wouldn't be trivial and would likely rely on consensus between community mirrors of the data, but the recovery path at least exists.

Cratesio package owners being (almost[1]) exclusively tied to GitHub auth services is a much more significant vendor lock-in than any other part of the Rust project's web presence. The cratesio team knows this and would prefer to not be thus reliant, but there's difficult work (both design and implementation) involved in accepting a second auth provider, and merely keeping the service running decently well keeps the team at (or beyond) capacity.

As with many things in Open Source governance, the status quo is good enough that nobody is particularly interested in doing the low impact dirty work to improve matters here. If someone (already reasonably reputable to the Rust project and with relevant domain knowledge) were to pledge a couple part-time person-months to an improvement, I would hope the team to be able to find enough time to review the work, since they've expressed some interest in improving the status quo in the area.


  1. There's the login tokens used for CLI authentication, of course, and those could potentially serve as an avenue for disaster recovery of accounts. ↩ī¸Ž

8 Likes

Thanks! I do have that same communication problem at work too. Drives them crazy.

2 Likes

This topic was automatically closed 90 days after the last reply. We invite you to open a new topic if you have further questions or comments.