„Ohne Helm und nicht angegurtet
Dazu komme, dass jedes zweite Kind im Lastenfahrrad keinen Helm trage und ein Drittel gar nicht oder nicht korrekt angegurtet sei.“
(Fragt mal so einen durchschnittlichen Niederländer, ob er nicht lieber mal mit Helm…?? Das Gelächter wird dich noch im Traum verfolgen.)
Traurig #WDR , sehr traurig.
For the past few years (basically within the first few months of using Emacs in 2020) I’ve been using #Helm. My understanding of how completing-read and the minibuffer in #emacs should work are thereby strongly influenced by this experience.
This weekend I finally had some time to play with #Vertico (and marginalia, consult, orderless… inspired by Prot's video). Probably I’m holding it wrong but it was not a smooth experience. Is it correct that vertico ignores the sorting of the collection used? Is there a way to change this? Why is the minibuffer here broken? Any advice would be welcome.
The images attached show the results for (completing-read "" orgrr-selection-list) in helm and vertico.
This is why I love #Renovate over #Dependabot: it picked up #Docker tags somewhere deep inside some #Helm chart, plus picked up all the Helm charts in my #TerraForm that need updating. Looks like I'll be updating everything tomorrow (well today since it's past midnight) 🎉🎉🎉#
Hi all, I've just migrated from @irfan to this account on sakurajima.social. I'll also be attempting to migrate my main account @irfan to this one in the coming days, but I'm expecting nothing on that front since it's failed so far.
I've moved to multiple servers throughout my time on the #Fediverse in pursuit for a "permanent" instance to be on, and I've never had the chance to write an #intro post lol so this will be my first.
I'm a 26 year old (in Nov 2024) (gfdi) programmer obsessed with languages such as #Python and #Shell, frameworks like #Django, and #CloudNative shenanigans like #Docker, #Kubernetes, and #Helm. I'm also obsessed with #FOSS and #Linux, and promoting them to be more viable each day for the masses so everyone could have not just free but secure access to daily computing.
I'll be using this account to mostly document my passions through mini reviews or rants. I have another picture-based account over on #Pixelfed, @irfan. All images I have posted and will be posting will include descriptive/alt text, including my pfp and banner. I also have some bots for posting local (to me, #Malaysia) #news and other relevant data cos the #Mastodon API is just so fun to play with.
A lot of my apps require a scheduler, I've always given the option to choose between using #APScheduler or #Celery/#Redis for that. So tempted to drop Celery though because of just HOW MUCH RESOURCES it needs to even work, and when it doesn't work it's all silent about it. In my #helm charts, I could set very low (default) resource limits because they've been tested to work/be sufficient with APScheduler and the rest of the app - because of Celery though, I'd need to set a much higher default.
I feel like I'm bombarded with things I wanna do devel-wise so I'm just gonna write them down here so I don't feel overwhelmed lol:
Work on my #Mastodon - #Djangoframework creatively named as mango to be used as a base for all of my other Masto-Django projects.
Port my other creatively named projects such as rizz (RSS-Masto bots like @lowyat and @soyacincau) and waktusolat (@waktusolat) to use mango as a base when it's finished.
New project based on mango to track/update local COVID-19 data on Mastodon.
Update my long abandoned clog (web blog/todo/money tracker app) project.
Develop a new web app, that could potentially be used as an e-invitation card/management portal. If this one goes well, I could self host/deploy it for my own wedding eventually lmao.
Create a Helm chart for #Kutt (FOSS URL shortener) and deploy it to my homelab.
Lots to do, for someone with extremely short attention span.
Is there really no way to programmatically create a #Helm values.yaml with our values specified through set etc.? Seems like a pretty simple use case they could handle but I'm not seeing any docus/discussions saying how or if we could even do this.
it me, mr reinvent the wheel cos i have tunnel vision every time ive a(n actual or artificial) problem to solve.
15 #helm charts to my name now, 2 of which in this list have been deprecated (postgres-agent replacing postgres-createdb and postgres-dropdb), 6 of my own (containerised) projects, and 7 of projects that either do not have existing helm charts of their own, or I just don't prefer/trust using them.
Een #helm op de #fiets is suggereren dat fietsen onveilig is. Wat veel beter is is investeren in zoveel mogelijk infrastructuur gericht op de fiets zelf.
Bij de zuiderburen is hier te weinig van. Fluohesjes en helmen zijn dan een weinig effectieve "oplossing".
"Am späten Mittwochabend verlor ein 24-jähriger #Raser die Kontrolle über seinen Wagen. Er rammte ein an einer Kreuzung stehendes #Auto und beschädigte drei weitere. Der 24-Jährige erlitt eine Kopfverletzung und kam zur stationären Behandlung in ein Krankenhaus. Es gab keine weiteren Verletzten. Der Autofahrer trug nach Polizeiangaben keinen #Helm."
<gripe>Why aren't postgres major version upgrades automagic in #Kubernetes but instead there's a whole tedious disruptive process? Why can't this just be solved with an init-container that runs pg_upgrade? How is this still a thing?</gripe>
If I had a very generous sponsor to cover the huge time commitment required for this, I would fork an #ActivityPub server (probably #Mastodon or #Firefish) and:
Rearchitect it to use industry-standard products for queueing, streaming, relational, key-value, and in-memory stores. Ideally everything must be so widely used that a major cloud provider offers a managed version of it.
Standardize the backend on one (1) technology wherever possible. Probably Python.
Making it as easy as possible to deploy on your own infrastructure or leverage managed cloud services. Also easy for the wider community to contribute. That's the dream!
Since quay.io is down, now is a good time to remind tech folks of reimage. You can use it as a helm post-renderer, it'll copy images from the random wilderness into your own OCI registry, and avoid runtime dependence on the likes of quay and dockerhub
Nice to see an official #helm chart for #Plex Media Server now exists, even if support for it is only community based. I also learned about rclone as a way to mount storage and make it accessible, though I question how performant it could possibly be.
reimage is a tool for #kubernetes and #helm/#helmfile that makes it easier to ensure your kubernetes deployments are more consistant by
copying images from remote registries to your own registry
updating image fields to a configurable consistent naming scheme, and optionally fixing them to digest forms.
. It also makes using Google Cloud vulnerability scanning and binauth attestation super easy.