💾 Archived View for gemini-textboard.fgaz.me › thread › 3 captured on 2020-09-24 at 00:40:15. Gemini links have been rewritten to link to archived content
-=-=-=-=-=-=-
Test thread. Put your test posts here so we don't pollute the main page
I can ad a post, but I'm anon... does this use a client cert?
trying another post, but with a client cert active. Will this work?
No cauth yet, sorry! -fgaz
auth* :)
does it works??? (this is neat btw tbh ngl; i've always thought the structure of imageboards was good even if literally every one is a nazi cesspool. better than opaque algorithmy nonsense a la reddit or w/e. sorting by activity is the best imo)
through the portal?
I also think the same, the structure is what's really great about *boards. Plus it's easy to implement, so it was a great first project
Oho, I see that someone is trying to mess with nonces already :)
To whoever tried to post "Everyone seems to be called anonymous", there's a redirect but with either my server or your client that makes it send the post to the wrong url
And yes, everyone is "anonymous" 'cause I have yet to add client cert support
s/redirect but/redirect bug/
Are new lines allowed? This should answer that question.
The answer appears to be: no.
how about [92mANSI escape codes?
I thought ANSI escape codes weren't part of the gemini spec?
Wow, a message board on Gemini? Cool! - Samsai
@32 not explicitly, but many terminal clients pass them through to the terminal, so they work. e.g. astrobotany uses this. I'm not sure it's a good idea, but there you go...
baschdel was here
This is annoying with a cache, I should really put a sideeffect database on my todolist
=> gemini://gemini-textboard.fgaz.me Do liks get escaped?
``` preformatted blocks?
> quotes?
Good job! Is there a feataure planned to post links?
Everything is escaped :) I built a library for parsing/printing text/gemini (will be published with the rest)
I did think about enabling links and other stuff, but left it as TODO. I'll open an issue about that. Ideally there'd also be a way to signal a newline, since the spec doesn't allow them in the input
Oh, and about escaping, I wish there was an official way to escape stuff. The backslash is kina ugly but at least it works
The spec doesn't disallow new lines in input at all, they just need to be properly percent encoded. The problem is mostly a UI one, i.e. most clients read only a line of input and stop when you hit enter. But there's no reason they couldn't provide some way to e.g. launch an editor instead.
As for escaping, I would strongly recommend making it clear that comments in applications like this are plain text and stick everything between preformat toggling lines. This does mean you'll need to wrap comments to a fixed width yourself, which might make them hard to read on mobiles, unless you wrap to 40 chars or so.
Using preformatted text would be a step backwards. I'd rather allow some line types
@44 yes, the spec does disallow newlines, if I unerstand this sentence correctly: "The requested resource accepts a *line* of textual user input"
@43 yes, there really should be an "escaped line" linetype. Having a backslash as the magic character to introduce it sounds good to me. It should work within a preformatted block, so you can have a line beginning with '```' within a preformatted black.
I don't understand the use of an escaped line, honestly. That much meta-discourse isn't necessary afaict
Well, it's not really necessary for self-contained lines... but it absolutely is for ```, which affects later lines too. Of course, adding another ``` on the next line should work around the issue, but isn't that nice imo
agreed that a way of sending multiple lines would be nice. have you given any thought to using the titan protocol for posting, if only optionally? it would seem to handle this problem while still being gemini-y: https://communitywiki.org/wiki/Titan
such neat
test from kristall! yoo dudes