Hacker Timesnew | past | comments | ask | show | jobs | submitlogin

As far as their code search under the hood it's just Elasticsearch, nothing special. Github's explanation sounds like mostly bullshit. Forced authentication of public endpoints is not an appropriate solution for the issue they're claiming. People building bots who actually care about this endpoint can just have their bots login with free accounts. This doesn't actually prevent load on their servers in a meaningful way.


GitHub's code search is no longer using Elasticsearch, it is using an entirely in house (Rust-based) search engine nicknamed Blackbird: https://github.blog/2023-02-06-the-technology-behind-githubs...


When they’re logged in, they are much easier to throttle or even ban compared to IP hopping to dodge IP throttling/bans.


>bots login with free accounts

those accounts still need to be created, which is hard to automate

also, now GH can just ban more easily detectable accounts that scrape, vs anonymous visitors




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: