I am pretty happy to see my patch for RubyGems (edited: Actually, it's Bundler!) from 4.5 years ago got merged: https://github.com/ruby/rubygems/pull/4892
Thanks @hsbt for taking care of that.
| Profile | https://www.godfat.org |
| GitLab | https://gitlab.com/godfat |
| GitHub | https://github.com/godfat |
| Taiwanese |
I am pretty happy to see my patch for RubyGems (edited: Actually, it's Bundler!) from 4.5 years ago got merged: https://github.com/ruby/rubygems/pull/4892
Thanks @hsbt for taking care of that.
Oh! I only realized now that my patch for rspec-mocks was merged 5 years ago: https://github.com/rspec/rspec-mocks/pull/1218
It's still in 4.0 beta though, no stable release yet 😬 https://github.com/rspec/rspec/blob/v4.0.0.beta1/rspec-mocks/Changelog.md#400beta1--2026-02-18
I understand it's considered a breaking change because it's fixing a "bug" that does change the behavior.
I am still pretty proud that most of the bugs I found in rspec mocks could not be reproduced in muack: https://github.com/godfat/muack
It's also pretty small, too! However it's fully feature complete.
@draconacht Yeah links are fake so it's not real. That makes sense.
As for trust, I think that's too heavy to put it. Always be skeptical. However I still believe we should encourage people to be honest. It's not black and white after all. You can call that being good or convincing, too. Just don't make it look like having obvious lies are ok even when it's common.
Maybe I should move to a Mastodon instance where I can post more than 500 characters. This limit annoys me to no end for English.
Chinese on the other hand, 500 characters can possibly contain 10x more information within the same limit... much less of an issue to be honest. This speaks that "characters" limit is quite arbitrary 🤷 Systems like this usually use "bytes" limit which in some way makes better sense after all. A Han character is 3 bytes in UTF-8, which is more "balanced".
RE: https://hachyderm.io/@rradczewski/116212327811508257
I didn't read through everything because it's quite long, and this is certainly quite mind blowing, yet it reads to me more like a human problem than software problem to be honest.
Reports were ignored, and people who actually used the software weren't trusted.
> Subpostmasters began reporting balancing errors within weeks of the Horizon system being installed, via the helpline they were instructed to use.
> ...