3 ms·
Its also a perf / ddos thing. If anyone from anywhere can run expensive searches it could degrade performance. A little log in won't hurt.
by blinded 3y ago
Its also a perf / ddos thing. If anyone from anywhere can run expensive searches it could degrade performance. A little log in won't hurt.
- tedunangst 3y agoNow I just clone the repo and run grep myself. Anything I can do to help lighten the load.
- blinded 3y agoSame, but silver searcher or fuzzy search ftw :)
- dumbo-octopus 3y agoFWIW, github.dev has a search feature that automates this flow using the fastest available subroutines for each step, all you do is a repo full text search with the same UI as normal VS Code. In my testing it searches about as fast as AG, and much faster than Grep. RipGrep beats it in a direct search time comparison, but the overhead of cloning the files is dominant in large repos. GitHub.dev searches via a .tar.gz archive stored in IndexedDB, which is much faster than cloning even at depth 1 and building up the files on disk (~50% speedup on my testing in very large repos, ~90k files).
- blinded 3y agoThanks for the pointer.