New cool feature on this morning: pinned tweets!

They show as native pinned toots and they are not affected by the 100 top tweets limitation.

To fix the crawling, I'm going to need to start using Twitter accounts to fetch new posts. This is getting complicated for the time I can put on this as a side project, so I'm going to rely on Nitter implementing this, then I'm going to use that. You can track progress here:

There is another thing I can fix natively in though. I noticed pinned tweets are there in user lookups. I can forward this, and the top 100 limitation doesn't apply to that.


Show thread

I'm getting a lot of question about what's currently working with, so here it is:

The system is mostly back online, but there are some major limitation when crawling new posts.

- To be forwarded, a post has to be in the top 100 of an account as measured by likes
- There are no retweets being forwarded right now


One really useful thing Nitter could do is to log the id of new tweets (and their author) as it encounters them during normal operation. could simply hook into that byproduct and it can already figure out the rest on its own.

I’ll sleep on it and if I still like it I’ll propose it to the nitter dev

And no retweets, but I did notice I can get pinned tweets though

Show thread

And we are back*, baby!

* with the limitation on popular tweets for now

Show thread

Currently setting up a mastodon instance and it looks like all requests have a 20 second delay added. It's driving me crazy!

Anyone knows what's up? My google-fu is giving me nothing

So they are all green right now, and the fetching tweets part works from my computer and CI, but not from any of the servers... The approach of fetching tweets from the embeds doesn't seem to work....

FML, and back to the drawing board!

Show thread

I will continue with something homegrown for users and individual tweets, but I'm thinking if I should add a dependency on nitter to fetch timelines 🤔

The lead dev has timelines working again already:

While having everything built-in is neat, strategies to crawl are getting more complicated and it's starting to take a lot of time to re-implement things in C#

Show thread

@vincent I'm loving the conflicting poll responses so far. "Fixing bird isn't worth it! But you should do it!"

I have also been prototyping for Instagram, which will be called

Should I focus my energy on that instead?

@jerome I designed from the beginning to rely on endpoints that would break major use-cases of their product like yours if they ever remove them.

Turns out Elon is perfectly willing to break them, which completely blows my mind


"a fascinating way to show users the refresh rate of their screen by spinning a specially designed zoetrope-like disc graphic at speed. As it spins, the numbers representing the viewer’s refresh rate should remain visually stable unlike surrounding numbers."

Looks like Twitter has made it so that when you look up a profile without being logged in, you see posts ranked by likes instead of by time. This will be a problem…

TweetDeck still runs, but it is hidden 

Guys I didn’t change anything yet and started working again 🤯

Show more
Librem Social

Librem Social is an opt-in public network. Messages are shared under Creative Commons BY-SA 4.0 license terms. Policy.

Stay safe. Please abide by our code of conduct.

(Source code)

image/svg+xml Librem Chat image/svg+xml