@bender@twtxt.net Ita disgusting 🤮 I can’t read shit 🤣

@slashdot@feeds.twtxt.net AI not living up to its hype?! Shock! Horror! 😱🤣 #AI

wut da fuq is this?! 🤣

@bender@twtxt.net I have not hmmm 🤔

@xuu Haha 🤣

@shreyan@twtxt.net Good morning! 🥱

@bender@twtxt.net Hah! 🤣

Good points 🙇‍♂️

@movq@www.uninformativ.de Yeah your original idea of precent encoding some information about the new follower is probably what we need to think about more. I think it’ll also work for Gopher/Gemini folk too right? So essentially new metadata key (optional) with some spec for encoding information about the new follower if either a) You don’t implement the User-Agent part of the spec or extensions or b)You use a protocol that makes this impossible.

@movq@www.uninformativ.de Sad, the search engine doesn’t have the full conversation 😢 I think I need to teach yarns how to crawl and index archived feeds 🤔

For HTTP WebSub is a good simple option here and there is this free inline WebSub hub anyone can use.

@movq@www.uninformativ.de Something like that, yeah 🤔

Similarly an optional subscription endpoint so we can optionally avoid having to pull feeds.

@movq@www.uninformativ.de I’m thinking of something like an optional metadata key that could be used as a general motivation endpoint?

@bender@twtxt.net thing I am too (happy) that is. If I ever wanted to side load an app I could with my Developer account easily!

@movq@www.uninformativ.de Yeah okay I didn’t notice that 🤦‍♂️ Kind of makes me wonder whether we should do something for the Gopher and Gemini folk here? 🤔

@movq@www.uninformativ.de Hmm wow! 😱

@aelaraji@aelaraji.com Well yeah this is true, but as @bender@twtxt.net and I point out, there are reasons for why this happens at times 🤣 Some of those reasons are not “purposefully malicious” I don’t think 🤔

@aelaraji@aelaraji.com Correct me if I’m wrong, but isn’t the EU forcing Apple to support side-loading Apps now on iOS devices? 🤔

Anyone had any intereractions with @cuaxolotl@sunshinegardens.org yet? Or are they using a client that doesn’t know how to detect clients following them properly? Hmmm 🧐

@falsifian@www.falsifian.org No need to apologize! 😌 Yarn / Twtzt is designed to be slow on purpose 🤣

@movq@www.uninformativ.de Where was this found? 🤔 What a treasure trove 😅 #Windows #sucks

@lyse@lyse.isobeef.org Poor little guy 😢 Also eww I hate snakes 🤣 I have a bit of a phobia 😅

@bender@twtxt.net That’s also likely 😅

@slashdot@feeds.twtxt.net IMO the problem isn’t “Apple”™ the company, per se, it’s the complexity and size of the organization. It’s large, it’s complex, you will have large sets of “people” that follow the “rules” as written without any leeway whatsoever or ability to apply “intent”.

@quark@ferengi.one probably not with the OpenSSH tools themselves but perhaps with the OpenSSL tools using the same key.

@off_grid_living@twtxt.net This is so true! Grandpa used to say this all the time too. Especially about tomatoes 🤣

@movq@www.uninformativ.de Yeah we’re about 3,054.26km from the equator but clearly that’s not far enough 🤣

@bender@twtxt.net Yeah just in spring right now and it’s already disgustingly hot 🥵

Uggh it’s 33C right now and 60% humidity 🥵

@quark@ferengi.one Bahahahaha 🤣

That was the only real way I could fix this for the time being without re-factoring the whole UI for the endpoint/page.

@abucci@anthony.buc.ci Precisely 😅

It’s a really good time to invest in nVIDIA shares 🤣

@abucci@anthony.buc.ci No this correctly now responds with:

$ curl 'https://twtxt.net/external?uri=https://google.com&nick=lovetocode999'
Feed Not Found

After nuking that from my cache. I forgot to nuke my own cache myself because it’s quite destructive and takes a few mins to rebuild on my pod 🤣 There’s a tool in tools for deleting a specific feed from the cache that I’ve been using.

@cuaxolotl@sunshinegardens.org Hey! 👋

Yeah I wonder too 🤔

@abucci@anthony.buc.ci appreciate it if you find the time to update again 🙏

Now that’s rolling out, I think that’s it. The only final way I can improve that /external endpoint/view is to refactor how it works a bit and add some HTMX magic™ so it has a nice snappy UX to it as it dynamically tries to validate the feed and provide useful feedback to the user, that way I can avoid injecting it into the cache unnecessarily in the first place!

@bender@twtxt.net Yeah but I found another bug and just squished that. CD pipeline is gonna roll this pod soon™ – Basically wasn’t handling feeds that redirect properly. e.g: https://google.com => https://www.google.com (though it’s not a feed 🤣)

Hmm I see this in the cache again 🤦‍♂️ Not sure how tbh – Job for me later.

@bender@twtxt.net https://google.com has been removed from the cache (without nuking the entire cache) @abucci@anthony.buc.ci if you need to selectively do this for some reason, there’s a script in the tools directory for this:

$ ./tools/cache_delete_feed.sh 'https://google.com'

Anyway, that’s gone. This is much much harder to exploit now, even if you’re an authenticated user.

Time for work™, But I quickly hacked together a bit of a better solution here. Rolling it out to my pod so we’ll see how it actually goes. Still possible to abuse if you’re a logged in user, etc, but at least now we delete the invalid/bad feed afterwards if it a) was not even a text//plain content-type or b) it errored out and was a new fetch of a HTTP feed.

Yeah okay.

@bender@twtxt.net Not possible 🤣

@abucci@anthony.buc.ci I mean it’s only suppose to do one thing really. What are you thinking here?

Yeah I was afraid of this. Technically can still be abused by “logged in” users. Hmmm

So we really not trust ourselves? 🤣🙄

@abucci@anthony.buc.ci No worries! 😅

But this is super weird, should behave the same as my pod 🤦‍♂️

@abucci@anthony.buc.ci nuke the cache file before starring