bowersbros

I've ran into so many problems trying to get a nice setup for my local development that matches similar to our production setup. I usually have completely separate docker compose setup locally, and a kubernetes setup of some kind in production.

These behave fairly similar, until I need to work on the ingress, or network policies, or anything else that is kubernetes specific.

I've finally taken the plunge into running it all locally on k3d, and have a pretty nice setup now with the manifests syncing with ArgoCD.

I'm thinking of packaging it up into a justfile and documenting it, but wondered how other people work on these sort of things. Do you just accept the shortcomings and test in staging envs for kubernetes specific changes?

It's a British colloquial term meaning basic / expected / ordinary

One reason would be that Vercel does the build process, whereas Github pages (afaik) would require you to commit the dist dir

You could already do that by setting a _acme-challenge alias CNAME record, so you can point it to a domain that they control. It's a bit finnicky and the new setup definitely looks better for whitelabelling, but there are current workarounds; although it does still involve client DNS access, you can put it onto a zone that has no risks

I had a teacher who told us to make a website using Powerpoint..

Turns out you save save as HTML and any links you put between slides become anchor tags.

Pretty neat, but hurt my soul to have all my classmates do that

Any chance you could share those sections of your claude file? I've been using Claude a bit lately but mostly with manual changes, not got much in the way of the claude file yet and interested in how to improve it

That first article is a decade old.

Only a few of these i found annoying, but its good to get a fix for the Previews and GitHub-Style Inline Diffs

But _most_ of the time you'd be using the one in your city.

If you are from Manchester, UK and you live there, how often do you need a taxi app that works in Bangkok, or NY, or Oslo?

Old School Runescape is where you want to be

Things like large documents don't make that any easier though. Scrolling through a large document can take a while, but a scrollbar allows you to get to somewhere else often quite a lot faster.

Imagine you cmd + F a phrase and there are 200 responses separated into like 4 chunks of a document. It can be useful to scroll to the start of each chunk to get the context of them.

To be fair, he works for Microsoft, its entirely possible he maintains one of their OS repos.

I can't see anything about the 33.0.0.0/8 range being reserved

https://en.wikipedia.org/wiki/Reserved_IP_addresses

Note that EU customers of Stripe get a drastically nicer 1.4% + 20p charge.

A better solution would be to use the status bar at the top, similar to how it changes when "In call".

Have it change colour, and say "1 new Mail notification" etc, and if you tap that, you get the proper context notification show up.

This is commonly solved using Amazon S3 and single-access tokens I think?

Pretty hefty work involved in getting it reliable, would be nice for a "drop in" solution.

Haproxy supporting `.map` functionality is the biggest selling point to me.

It allows for SSL protected custom domains for a user of a SaaS app in a really easy to do way.

We use it at Shopblocks extensively, and I've written a (brief) outline of roughly how we do it.

https://zando.io/post/haproxy-saas-custom-domains/

The problem being solved there is being able to see further in the dark.

Those 25 countries are where to _seller_ can be based. Stripe allows purchases from over 135 countries.

https://stripe.com/docs/currencies#charge-currencies

A huge backwards incompatible change

As I understand it, repairs are expensive, but general maintenance is less, because there isn't as much to maintain.

One of the issues we had with wkhtmltopdf was that we use Angular for our views, and we have a view for an invoice, and the ability to print it.

We wanted the same view to be passed into wkhtmltopdf to download it as PDF (with a couple of things omitted).

Unfortunately, the JS didn't execute properly, so you get a borked output.

End result was we render it, capture the HTML after render, and pipe that through to wkhtmltopdf.

Using chrome _should_ resolve that issue and let us just pass the page through.

Do you have a mailing list where I can say my bank is with "Halifax let me know when you support them securely?"

I'd like something like this, but I wouldn't want to give the bank details away.

Any chance I can get in on the beta for this? I've done the Load balancer and block storage beta? <username>@gmail.com

because the artwork is of a nude 10 year old.

What is your emails bounce rates etc? Or is it caused by others using those same IPs too?

More money wouldn't help that cause though. The US doesn't need the billions to solve issues, it already has it. The lack of solutions in the US is due to people not working towards them, not the resources being lacked.

We outsourced our database to Google Cloud after having pains with percona on DO with clustering. Might be worth looking at that. No regrets for us so far, except that restoring backups is painful if you want 1 out of many databases restored.

We use Digital Ocean at shopblocks.

We use it for both production, and staging, though we do offload our database to Google Cloud.

On DO we host our application servers (frontend and backend), as well as our caching, search and monitoring services. These are all load balanced behind Haproxy.

I'll be writing a blog post in the coming week and publishing it on HN regarding how we do it; its a simpler method I think, though we aren't the scale of Etsy yet, but should be useful for smaller startups to adopt.