💾 Archived View for station.martinrue.com › elektito › tinylog captured on 2023-04-26 at 14:57:01. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2023-04-19)
-=-=-=-=-=-=-
author: elektito
Crisis averted!
Looks like I've broken my search engine without even knowing it...
So I just realized how much a trailing slash could matter in urls: if I go to my gemlog at gemini://elektito.com/gemlog (notice the lack of a trailing slash) and click on one of the relative links (like pointing to "1-first-post"), the client asks for gemini://elektito.com/1-first-post, instead of gemini://elektito.com/gemlog/1-first-post. At first I thought this is a bug in the clients, but now I'm thinking it sorta makes sense. Without the trailing slash, the browser asks for a sibling page, same as if I was looking at gemini://elektito.com/gemlog.gmi. I'd better reconfigure my server to redirect those urls.
So over the weekend I added "image search" to my search engine, "image" here meaning ascii art! You can try it here: gemini://gemplex.space/image/search.
There's also an option to show a random ascii art: gemini://gemplex.space/image/random
Is it just me being fixated on this, or have the web search engines become entirely useless? Whatever I search for, I seem to get the same 17 best this, and best alternatives to that, and 21 great...and so on? The only way I seem to find stuff is when I already more or less know where something is to be found, like reddit, stack overflow, wikipedia, you know the obvious stuff.
Funny thing about Station is that there's really no need to follow anyone atm. You can just check the front page! I like that.
I talked about moving away from services like github a couple days ago. Here's the page I've setup to track my progress. It's a little unweildly for now. Guess I'll have to split it up and organize it into multiple pages soon. gemini://elektito.com/project-migration
I see Google-style robots.txt rules like /*.jpg$ in the wild. As far as I understand, this is not valid under gemini robots.txt spec, or at least not like it's intended. Or did I get that wrong? Anybody else following those rules?
Let me reiterate in this venue that I _hate_ Ubuntu snaps. They never work quite well. For example, the way Firefox updates are handled is just abhorrent. You get a notification that there is an update pending, but Firefox needs a restart, but it won't really work unless you do that in the short window that the update is actually retried.
Just came across this post: gemini://ploum.be/2023-02-22-leaving-github.gmi. I just needed a gentle reminder to get out of github. I'll be making preperations to make the migration soon. I already have a paid sr.ht account, so that's where I'll be heading. I can also think of other services that I want to start migrating out of. I'll probably make a page on my capsule to keep track of what I need to do, and keep a log of things that are done.
I am letting my 2325 day Duolingo streak die. Duolingo quickly turned out to be not that educational to me. I just kept on the streak because it's the sort of thing my brain keeps forcing me to do. But at this point, the amount of streak freezes that I use is just rediculous and the whole thing is entirely pointless. Not to mention the fact that Duolingo has become complete enshittified as Cory Doctorow would put it.
Deciding it's never going to be "ready" ready, here's an early version of my gemini search engine: gemini://gemplex.space/
Finally added "slow down" support to the crawler, in case any server finds the default rate of one request per second too much. I'm a model citizen now.
On my server logs, I see requests to gemini:///robots.txt. Is this supposed to be a valid request? Should the server fill in the missing hostname with SNI maybe?
Everytime I hear "gemini is heavier than gopher, lighter than the web", it makes me think of "duty is heavier than a mountain, death is light as a feather."
Occsionally someone comes up with something so weird and so interestingly, that it absolves the rest of the internet for a while (only a short while!). You think you could use a sector on a real floppy disk? Grab one, while they last. There's only 2880 after all! https://sectordisk.pw/
Moved my capsule to my own Gemini server, which is now called "hodhod". Source on github: https://github.com/elektito/hodhod
Geminispace seems to be entirely devoid of anything related to the Penguins of Madagascar. That's a real shame! So I just added a page to my capsule, dedicated to collecting some of their quotes.
Added a couple of silly mountains to my capsule's ascii art header. I feel like quite an artist now!