Conversation

Notices

1. Niki Tonsky (nikitonsky@mastodon.online)'s status on Monday, 25-Nov-2024 09:45:14 JST Niki Tonsky

   - Run parser on file containing only ASCII, get X µs
   - Run parser on file containing wider UTF-8
   - Run parser _on the same first file_ containing only ASCII , get 2X µs now and from now on
   - Reloading code helps, until I do UTF-8 once again

   What’s going on?

   • Niki Tonsky (nikitonsky@mastodon.online)'s status on Monday, 25-Nov-2024 19:43:41 JST Niki Tonsky

     in reply to

     • Jack Rusher

     @jack the file looks like this. It’s all ASCII, it only decodes into Unicode

   • Jack Rusher (jack@berlin.social)'s status on Monday, 25-Nov-2024 19:43:42 JST Jack Rusher

     in reply to

     @nikitonsky I would be interested to see if this behavior changes when you put an encoding specifier on that middle slurp 🤔

   • Niki Tonsky (nikitonsky@mastodon.online)'s status on Tuesday, 26-Nov-2024 01:43:49 JST Niki Tonsky

     in reply to

     Update: by removing one line at a time, I figured that `let += nchars;` is responsible for the slowdown.

     What’s weird is that `append` function itself is called in both slow and fast cases, so it’s not like the function itself is slow, it just _becomes_ slow after working on certain file