Christian Speich

@kleinweby
9 Followers
37 Following
12 Posts
Show threadChristian SpeichSep 20, 2025
@343max das nächste MacBook kommt dann mit dem passenden Radius beim Bildschirm. Super Grund für ein Upgrade 😬
Christian SpeichAug 10, 2025
SaraMGJun 12, 2023

Hiring a software engineer in their late 40s:

Pros:
* Understands your stack better than you do after glancing through the repo for five minutes.
* Will rewrite said stack 2x as fast, and half as buggy if you let them.

Cons:
* Gives zero fucks.
* Knows we're not *really* like family here.
* No, seriously, absolutely zero fucks given.

Do not cite the deep magic to me, product manager, I was there when it was written.

Show threadChristian SpeichAug 7, 2025
@mjg59 I’ve no idea what the specs say, but in Germany (the EU?) emergency calls require a working SIM card.
Show threadChristian SpeichAug 1, 2025
@343max Also ich kann `registry.gitlab.com/fmd-foss/fmd-server:latest` einfach ohne token pullen. Und eingeloggt bin ich auch nicht.
Show threadChristian SpeichAug 1, 2025
@343max Personal kann das auch. Du brauchst den Token um das Docker Image zu pullen? Das sollte bei public ja eigentlich nicht notwendig sein? 🤔
Show threadChristian SpeichAug 1, 2025
@343max GitLab Tokens können sich selbst erneuern, was irgendwie Quatsch klingt von der sicherheitsebene her. Aber Achtung der alte Token wird beim selfrotate direkt ungültig: https://docs.gitlab.com/api/project_access_tokens/#self-rotate
Project access tokens API | GitLab Docs

GitLab product documentation.

Show threadChristian SpeichSep 3, 2024
@ivory Oh that’s great! Thanks 😊
Christian SpeichSep 2, 2024
@ivory wasn’t there a setting somewhere to auto-expand long toots in the “detail view”? Currently, I have to go deeper and deeper when reading a long toot with long follow-ups by the author :/
Christian SpeichDec 14, 2023
@[email protected] Did you see any error indicators (e.g. underun) on I2S in the error case?
Christian SpeichDec 14, 2023
@[email protected] oh what fun. I had a similar issue with mixed up channels on an iMX.8 once. Was quite fun to debug, however it turned out to be a software issue as NXP decided to make the submit dma call into a async work item. This meant that the dma was sometimes setup after starting the I2S and thus missing a few samples at the beginning, resulting in the misalignment.