Comment by 🚀 stack

=> Re: "After first using Cyrillic text in replies, found that..." | In: s/Gemini

At 4 bytes per char, it's still 256 chars... Is that really not enough?

As a last resort, keep the header in ASCII?

Oh, you mean for interactive input.

=> 🚀 stack

Jan 11 · 8 days ago

4 Later Comments ↓

=> 🦂 zzo38 · Jan 11 at 21:40:

I just use a separate program for Titan (so that I can store drafts in separate files). This does require that you can use the certificates with the separate program too, but that isn't so difficult to do, at least in my case.

=> 👻 ps [OP] · Jan 11 at 22:29:

The problem is my client (Yoda) does not support Titan yet, I plan to implement this feature, but don't understand why this Gemini limitation wanted for, just interesting where is benefit of 1024 bytes restriction.

=> 🚀 stack · Jan 11 at 23:24:

I think it's just stupid.

=> 🧇 Acidus · Jan 13 at 19:32:

🤷 There has to be some limit to the size of the URL or the size of the request line. Web browsers usually freak out on URLs beyond 2048 characters, though you could make the argument that Gemini URLs should be able to longer than 1024.

Since all user input comes in the form of a query string, and those values need to be URL encoded, the over all URL length limit ends up capping what can be sent, which kind of sucks, but it's not terrible. Station functioned for years without needing more. BBS has some interesting, though somewhat clunky work arounds

Original Post

=> 🌒 s/Gemini

After first using Cyrillic text in replies, found that default 1024 bytes in header definitively not enough, especially, when part of these bytes taken by path and auth token there. Seems it's time to implement Titan protocol in my app, but finally I don't understand where is rational goal to limit headers if the server able to make own limitations when that's wanted..

=> 💬 ps · 5 comments · Jan 11 · 8 days ago