Posts tagged: dev

All posts with the tag "dev"

285 posts latest post 2026-05-02
Publishing rhythm
Mar 2026 | 3 posts

import os import arel from fastapi import FastAPI, Request from fastapi.templating import Jinja2Templates app = FastAPI() templates = Jinja2Templates("templates") if _debug := os.getenv("DEBUG"): hot_reload = arel.HotReload(paths=[arel.Path(".")]) app.add_websocket_route("/hot-reload", route=hot_reload, name="hot-reload") app.add_event_handler("startup", hot_reload.startup) app.add_event_handler("shutdown", hot_reload.shutdown) templates.env.globals["DEBUG"] = _debug templates.env.globals["hot_reload"] = hot_reload @app.get("/") def index(request: Request): return templates.TemplateResponse("index.html", context={"request": request}) # run: # DEBUG=true uvicorn main:app --reload

I just discovered arel for hot reloading python applications when content changes from this snippet that implements it for fatapi.

...

why-is-postgres-default

Serious question.

But, why. It’s the most loved db, right? Right? Maybe it’s time to rethink it.

Don’t get me wrong, if I need a relational db as a service, PostgreSQL is going to be my first choice, but why do I need to run a separate application for it?

...

Pagefind is absolutely insane. I’ve tried a number of static site searches, and found them all hard to get get going, clunky and not the best experience as a user or developer.

I setup pagefind in about 2 minutes on my site where it found and indexed 833 pages in 2 minutes.

The only downside I see so far is that it is a lot of bandwidth to the user. On simulated slow 3G you can definitly feel it, but not terrible. Anything slower and its going to start feeling frustrating.

edit: I have actually fully deployed it on waylonwalker.com, and its fast!

...

A Case For Tailwindcss

I was watching @theprimeagen recently and I think he sold me on using tailwindcss. The thing about tailwind is that it is not a big component library, it’s a set of css classes mapped to a few (usually one) style.

All css classes are shitty, so you might as well use someone else’s shitty css classes on all your projects rather than thinking you’re being smart with a new set of classes that you will hate in 6 months when you come back to the project. roughly quoted from memory of @theprimeagen

So unlike big component libraries like tailwind, it comes with a cli that that it uses to create the final css file. It is able to treeshake out all the tailwind classes that you are not using and only ship the ones that you are using.

...

This is the greatest nvim emmet plugin I have tried. In the past I had tried the vim plugin a few times and just could not get a good flow with the keybindings and found it confusing for my occasional use. emmet-ls just uses lsp-completion, so its the same flow as other completions.

You can try it out by installing with :Mason

The htmx-request class is added to htmx-target elements. You can target this css selector to create loading state throbbers.

By default the target element will the self, but you can use the typical htmx css selector to select which element will recieve the htmx-request class while the request is running.

The only way to override the name of the class is through config.

Three ways to support updating other content. Fantastic article walking through the different ways to update other parts of the screen using htmx.

In htmx there is no 2 way data binding, the dom is your state, and if you have elements derived from the same data on the screen in different places you need to think about how to keep them in sync.

> A command-line utility for taking automated screenshots of websites

Daaaang, this is such an elegantly simple way to get web screenshots with a cli. I was literally up and running with two commands on my arch linux machine (which it warned was unsupported by playwright).

pip install shot-scraper # Now install the browser it needs: shot-scraper install shot-scraper waylonwalker.com shot-scraper https://datasette.io/ shot-scraper https://datasette.io/ -h 1280 -w 1920 shot-scraper https://datasette.io/ -h 480 -w 720 shot-scraper shot --selector '#posts' https://thoughts.waylonwalker.com/post/89

Note shot-scraper https://datasette.io/ is a full length screenshot of the entire page.

Oh and its pretty dang fast, let alone the setup time, this crushes on startup time in my attempts to use a headless browser in the past.

Most of the time when creating links in html you want to maintain the default behavior, as this is what users are going to expect, but sometimes your site behaves such that it does not fit, and it does something unexpected anyways. in this case you might want to make the default behavior to open the link in a new tab rather than relying on users to control click.

Use this with restraint as this can make your site feel janky and do things that do not feel natural to the web.

I went down the route of leveraging the json-enc extention in htmx, but later realized that this completely breaks browsers/users who do not wish to use javascript. While most of the web would feel quite broken with javascript disabled, I don’t want to contribute to that without good reason.

Taking a second look into this issue, rather than using json-enc, and using as_form to get form data into a model keeps the nice DX fo everything being a pydantic model, but the site still works without js. with js htmx kicks in, you get a spa like experience by loading partials onto the page, and without, you just get a full page reload.

copied from