This action will delete this post on this instance and on all federated instances, and it cannot be undone. Are you certain you want to delete this post?
This action will delete this post on this instance and on all federated instances, and it cannot be undone. Are you certain you want to delete this post?
This action will block this actor and hide all of their past and future posts. Are you certain you want to block this actor?
This action will block this object. Are you certain you want to block this object?
Are you sure you want to delete the OAuth client [Client Name]? This action cannot be undone and will revoke all access tokens for this client.
Are you sure you want to revoke the OAuth token [Token ID]? This action cannot be undone and will immediately revoke access for this token.
#crystallang 75 hashtags

Release v2.4.2 fixes a few more bugs. Only one is a regression—I found the others while testing. In this release:
Fixed
del, ins, and s elements in sanitized HTML.redirect_after_auth_path on browser navigation.FileUtils.mv to move uploaded files. (fixes #117)Thanks to @jayvii for help with troubleshooting the last one!

Eh, I didn't test enough and released v2.4.0 of Ktistec with a few annoying regressions. Release v2.4.1 fixes them! ☹️ 😠 😡 🤬
Fixed

I just released v2.4.0 of Ktistec. This release encompasses a few things that I've been working on for a while: improved support for operating without JavaScript available/enabled and support for running scripted automations.
Except for a few items, Ktistec now works without JavaScript. Obviously, things like WYSIWYG editing of HTML don't work—I plan to add support for Markdown to compensate. Running in Lynx is a stretch, but...

Since the early days, most controller actions supported both text/html and application/json. I cleaned up support for the latter and have officially documented the Ktistec API in the README.
In addition, I've added support for running bots/automations (prior announcement). The Ktistec server will periodically run any executable script in the etc/scripts directory. These scripts have access to the Ktistec API and can post, follow, share, like, etc. This is experimental and obviously introduces an attack surface, though that shouldn't be a problem on correctly configured hosts.
Here's the full changelog:
Added
Fixed
Changed
formaction. (fixes #101)Other

And we have a new release: Crystal 1.14.0 🚀
Read about the highlights in https://crystal-lang.org/2024/10/09/1.14.0-released/
Happy Crystalling!

have i told you all i love spectator? i love spectator!

I just released v2.3.0 of Ktistec. It fixes a few bugs introduced in the previous release. In particular, it handles a case I've observed a few times in production where the fetching of followed content hangs. The fix adds a monitor that periodically checks for "running" tasks without a backing fiber and sets socket timeouts to ensure requests eventually terminate (I believe the latter actually fixes the problem, but belts and suspenders...)
#ktistec #crystallanguage #crystallang #activitypub #fediverse

Today we celebrate the 12th anniversary of Crystal!
The initial commit was on September 4, 2012 🎉
https://github.com/crystal-lang/crystal/commit/220bb8c736381d5beda5e7bc36c8d7f0d59badf6

I just released v2.2.0 of Ktistec. A lot of code was written but the main user-visible change is streaming timeline updates to the web client using server-sent events. The back-end pub/sub framework will also make it easier to decouple some components down the road.
I want a visual indicator that new posts are in my timeline. I appreciate that not everyone feels the same way—if there's interest I may make it switchable.
And now, off to #worldcon!

the last few days are more in line with what i expect. there is a (temporary) bump in memory, but that corresponds to a large increase in traffic.


for reasons i don't understand, some restarts of epiktistes result in huge memory usage spikes—reported both by the garbage collector and the operating system. what's interesting is heap size and free memory move roughly in tandem (though the difference isn't constant), which implies that there is free memory, but perhaps it's too fragmented to be useful?

a restart typically fixes the problem...