At #SUSECon2026 it's all AI, Sovereignty and Resilience front and center, coloured in a hopeful green theme.
I got some new ideas already.
Remote
Ralf Lang
@ralflang@phpc.social
mastodon 4.7.3PHP Enthusiast. Former Perl Monk. Software Architect.
Housekeeping on fig/log-test package.
Somewhat composer/satis interested.
Blogs on all things Horde.
Likes trains and mountains and old castles.
13 Followers
34 Following
11 Posts
Joined October 31, 2022
Website:
Replying to
@dseguy@phpc.social Yes, that's the one. With pecl increasingly vague and pie coming but not yet fully arrived in everybody's mind it's sometimes hard to understand what is essentially continuation and what is a fork. I hope it's good contributions, I haven't done php c code in a while.
Horde.org 6.0 is in public beta. This week is dedicated to null safety and string conversion issues as well as chasing down user reported problems in ActiveSync support. We target PHP 8.1-8.5 for initial release and RC phase in May. Github issues are open :)
I am listening to the PHP Podcast
https://www.phparch.com/podcast/the-php-podcast-2026-04-09/
Quite entertaining.
Github is strange these days. Repeatedly it happened that it left PRs open after I pushed rebase-merge and then refused to redo/close "Base branch has changed".
Used to be kind of a nice product.
Horde 6 RC phase is set to start on 2nd of may and there's so much left to do. So few evenings left.
I won't do this very often but it gave me a chuckle. First addendum says - "Biking lane closed". Second addendum says - "This also applies to bikers".
My car repair shop has "upgraded" its hotline from some old school press this for topic A to a dumb ai bot including fake keystroke sounds and delays on every answer. It doesn't know several domain words (hardly exotic) and tries really hard to turn me off when I ask for a human operator. It's taken 10+minutes to get into the 1-minute call with a human who adresses my need. Give me back elevator music and hold the line loops.
At work in an AI heavy environment I sometimes let the agent just try to implement some spec not with the goal of actually using that but to use it as an input for critique, especially when it gets stuck and tries increasingly zany hotfixes. Learning from failure is a thing but learning from AI failure is a thing, too
Replying to