Jeff Triplett

Continuing in my tradition of 2023, 2022, 2018, 2017, and 2016 “DjangoCon US Talks I’d like to see” annual posts, here is my update for 2024.

This year’s updated talk and tutorial deadline is April 29, 2024, at 12 PM EDT, but check the official speaking page for updates and submit your proposal.

Before you read mine, I suggest you check out Kati Michel’s and Drew Winstel’s list of talk ideas too.

Modern Django

Three or four years ago, our Django stack was optimized to leverage the Django ORM to tie together a mix of authentication and REST API using the Django REST Framework. The front end was entirely buried by JavaScript, which printed HTML into your browser, and the HTML was entirely hidden away. What used to take a developer hours and days now took weeks because there were so many layers and abstractions to wade through.

Then it all started to change a few years ago, and we learned that we don’t need the additional weeks of development time to use JS/JSON layers for which we may never build that iOS/Android app. We don’t need React for contact us forms, and 90% of our web tasks could be built in a day.

The theme I’d like to see more of this year is doing more with HTML instead of hiding it behind layers of JavaScript.

I want to see lots of talks about technologies like HTMx and frameworks like Tailwind CSS, which are firmly centered inside HTML, instead of trying to avoid using HTML.

The two non-Python languages we are starting to talk about

Rust and WASM are starting to change everything. How do they fit in with Django and Python?

Theme: Rust

The Rust language is shaping the landscape of Python productivity in ways that none of us could have predicted. Tools written in Rust are significantly faster and are gaining steam.

• What can a Django developer take from Rust?
• How do I develop a package for Django and Python that uses Rust?
• How libraries like PyDantic are speeding up development (Django Ninja uses it)

Always nice to have talks

• What’s new in Django 5.x?
• What’s new in Python 3.13?
• What’s new in Wagtail CMS and Django CMS?
• Where do we want the DSF to go?
• How to contribute to Django or third party projects

Embracing HTML and lightweight front-end frameworks

Last year, I predicted it would be the year of “you probably don’t need JavaScript,” and I still feel good about that. We saw HTMx take off, but I think Django is still underserved and could benefit from more talks:

• Django 5 added more granular template/form styling , and we could use a talk that deep dives
• Using Django forms with htmx
• Using search with htmx
• Whatever cool things you are doing that I didn’t mention here

Panel ideas

• Django Fellows panel
• Django content creators' panel
• Django apps and community - It would be nice to hear from the more significant apps and community projects.

Wagtail CMS

• How do I migrate from WordPress to Wagtail CMS?
• How to properly test with Wagtail CMS for me, because I struggle with it

Talks I don’t want to see

I find myself at a weird crossroads because I have never discouraged certain topics, but we need to talk more openly about what is bad for our industry and the world instead of avoiding it. Otherwise, we are left with dead air, and people assume we are all on the same page.

To kick us off, there are talks and topics that I do not want to see:

• BAD AI talks and ones that spread more FUD into the community
• crypto talks
• why technology sucks
• proposing a talk because you dislike or even hate something
• bashing other communities
• bashing our communities
• lightning talks disguised as questions (not so much a talk idea, but this has been on my mind for a while)

So please, none of these:

• The future of Django depends on AI
• Why tech needs AI
• Crypto talks
• Your product unless you are sponsoring.

Good luck

You have less than a day, but between my list, Kati’s list, and Drew’s list you have no excuses for saying you didn’t know what to talk about at DjangoCon US this year.

Good luck!

Last year, I switched to the Vivaldi web browser because I read a few peer reviews that pointed out that Vivaldi is a better Google Chrome web browser than Chrome.

This week, Vivaldi rolled out a new memory-saver feature that I’m excited to try out. It lets you decide how long to keep a tab in memory before hibernating and reclaiming the memory dedicated to the tab.

Vivaldi’s memory and CPU usage are already night and day better than Google Chrome. For instance, I can have one Chrome window with three tabs open ironically for Google Gmail, Google Calendar, and Google Docs, which eat up more resources than 2 to 4 Vivaldi windows open with countless tabs.

As it turns out, all that tracking and spyware eats up a ton of RAM and CPU. Who knew? 🤷

All of my Chrome Extensions work on Vivaldi using the Chrome Store except for my custom Django News “newtab” extension that I wrote. I suspect it’s me, and I probably need 15 minutes to troubleshoot it.

Instead, I’m nudging you to drop Google Chrome and try out Vivaldi because several of my Google friends were fired this week despite the company never being more profitable.

ChatGPT told me that Alphabet/Google are worth enough to give everyone on Earth $253.16, which feels like a fair tradeoff. I’d like mine to be a check because I do not trust Google with my bank details, nor should you.

What I don’t like about Vivaldi

Vivaldi is outstanding and could be better.

The first thing you’ll want to do is access the “Settings » Address Bar » Drop-Down Menu Priority” and uncheck the “Direct Match” option.

The Direct Match option tends to match from “Brand websites,” which is one of the ways Vivaldi makes money, but it’s not the only way. When Direct Match is checked, I get recommendations for Macy’s instead of Mastodon until I discovered how to turn it off.

Check out What’s Vivaldi’s business model? to better understand how they work.

I would love to see Vivaldi let me PAY THEM MONEY annually for a solid web browser that aligns with my interests. Until then, I’m OK with them getting an affiliate cut from eBay or Amazon when I use their bookmarks.

Why not Firefox?

I hear this often, and I won’t talk anyone down from their favorite browser unless it’s Chrome or whatever Microsoft calls their latest hobby browser this decade. 🤷

Firefox doesn’t work well for me, and I say this because I have used it for over a decade. I also don’t trust their parent company because of where their money comes from and the various mistakes they continue to make.

If Firefox sparks joy, I encourage you to use it. I continue to get “the emperor has no clothes” vibes from their company, even when I want to like their products.

Share your thoughts

If you have thoughts, I’m on Mastodon, and I’d love to know what config options worked best for you.

#Enshittification

At night, when I’m couch computing after everyone goes to bed, I use one of these XL adjustable lap desks with my MacBook Pro (2019 model) and a cheap, wired external mouse that I love using. camelcamelcamel.com/product/B… This desk is wonderful, and I highly recommend it. 😘

However, while it’s the perfect height/location for typing, the built-in trackpad is more problematic to use to the point I don’t use it.

So, with these factors in mind, I’d like to know if anyone has an external Trackball they want that would pair well with this setup. I’m open to all suggestions and eager to hear your experiences. 

Historically, I have had fantastic luck with Logitech keyboards and various products over the years, but I’m a little put off by their recent “Put an AI on It” trend. So I’m not against it, but I’m not excited about it. 

Do you have any recommendations for a good trackball that’s not Apple’s Magic Trackpad?

My first PC as a kid ran MS-DOS 4, so seeing today’s news on Open Sourcing DOS 4 takes me back.

Today, in partnership with IBM and in the spirit of open innovation, we’re releasing the source code to MS-DOS 4.00 under the MIT license. There’s a somewhat complex and fascinating history behind the 4.0 versions of DOS, as Microsoft partnered with IBM for portions of the code but also created a branch of DOS called Multitasking DOS that did not see a wide release.

I played with LOGO and Applesoft BASIC on my grade school’s Apple IIe, but GW-BASIC shipped with MS-DOS 4 and was the first programming language that I knew inside and out.

I struggled at times because book and magazine examples were written for BASICA, which mostly worked, but it sometimes took some extra fiddling.

MS-DOS 5.0 exposed me to QBasic, which was the first programming language I worked with that had an IDE.

I fondly remember writing code on that old 386 and playing DOS games. I remember thinking there would eventually be no limits between the code and what I saw was possible on Star Trek: The Next Generation just a few years later.

Looking back, we aren’t there yet, but we are closer to the Star Trek computer and Holodeck than to the technology I grew up in.

Please note: I moved the time up 30 minutes.

📅 I’m hosting Office Hours this Friday, April 26th, 2024, from 2:00 PM to 4:30 PM Central Time (US and Canada). If you need to convert it to your local time zone, use this handy link.

ℹ️ During our session, we will discuss our ongoing projects, catch up with one another, and work together to wrap up our week on a productive note.

🙏 As always, everyone is welcome to join, whether you’re a regular attendee or joining for the first time.

✅ If you need any additional details, feel free to send me a message or check out the gist from our previous sessions, where you’ll find the Zoom link ⚠️

I look forward to seeing everyone.

If you use GitHub Actions to build and deploy Docker container images, you quickly learn that deleting old container images you no longer need is a huge pain. I don’t know why GitHub doesn’t make this easier, but they have a reusable workflow we can use to help prune and clean up old container images.

The delete-package-versions will let you set a number to only keep the last x-known good images, with some other valuable options like deleting untagged container images.

I prefer to write one reusable workflow, which I can call from my standard CI build and deploy pipeline, along with a workflow_dispatch action we can call whenever we want.

This workflow is a mouthful but wraps the delete-package-versions action and gives us a few configurable options to pass. Having a default value on min_versions_to_keep is nice because it is a default when calling this workflow from other locations.

# .github/workflows/delete-container-versions.yml
name: Delete Untagged Container Versions

on:
  workflow_call:
    inputs:
      package_name:
        required: true
        type: string
      package_type:
        required: true
        type: string
      min_versions_to_keep:
        required: false
        type: number
        default: 30

jobs:
  delete_versions:
    runs-on: ubuntu-latest
    steps:
      - name: Delete all untagged container versions
        uses: actions/delete-package-versions@v5
        with:
          delete-only-untagged-versions: 'false'
          min-versions-to-keep: ${{ inputs.min_versions_to_keep }}
          package-name: ${{ inputs.package_name }}
          package-type: ${{ inputs.package_type }}

Our cleanup.yml is set up as a workflow_dispatch, allowing us to call our delete-container-versions job whenever we want.

# .github/workflows/cleanup.yml
name: Delete Untagged Container Versions

on:
  workflow_dispatch:

jobs:
  delete-container-versions:
    uses: .github/workflows/delete-container-versions.yml
    with:
      package_name: 'your-docker-container-package-name'
      package_type: 'container'

I’m embarrassed to admit I had several repos with over 500 or more Docker container images. Deleting the oldest Docker images took less than a minute once I had the button set up.

After successfully building and pushing a new container image to our registry, I also include this job as the last step. This will future-proof our build and deployment workflow so that it never gets out of control again.

Also, I can count the number of times that I have needed to revert to one image on one hand and have never reverted to an image more than that. So saving the last 10, 20, or 30 images is peace of mind.

Quick update tonight since I’m tired, but I had to re-do/re-send Django News #228 because our newsletter provider failed, and feel free to read the rest.

Then I typoed DjangoCon US’s WEDNESDAY CFP deadline, which I fixed online, but I can’t fix what I already sent out.

I’m quitting early tonight.

The Nintendo Switch is very popular in my household. My six-year-old is at the right age to play a round or two of Mario Kart, TMNT, Kirby, or whatever we feel like, and it’s a fun night. What gets annoying is taking the Switch out from under our TV to switch games and wait for it to reconnect to the TV.

I was sharing this “pain” with a friend of mine who shared this Unitek 4-in-1 Game Card Reader (similar model) with me, which might be the most lazy product that I own. The Unitek reader connects to your Switch and holds four games, which you can switch between with a remote button. The product has no frills, but it doesn’t have to. All it needs to do is switch games when you press a button, which it does wonderfully.

Oh, a whim, I decided to check out their website to see what other products they make, and I stumbled on their Game KVM Docking Station with Remote. If I needed a Dock, I would be tempted. TBH, I’m already tempted to know if I could stack my 4-in-1 Game Card reader with the KVM Docking station to get seven games on standby, but I suspect the remotes will be on the same frequency.

If video game emulation is your thing, I picked up Delta last week for iOS in the App Store, but I have yet to have time to try it out. I own several small portable emulation devices that I rarely use, but Delta might be the one that finally sticks because I always have my iPhone with me.

Hug your loved ones. Everyone should take time and appreciate where we are and those around us on our journeys. We should also remind ourselves of everyone who has joined us along the way and what they have meant to us. 

To my friends struggling with various degrees of bad news today, I see you and wish it weren’t the case. ❤️

In March, I wrote about my robots.txt research and how I started proactively and defensively blocking AI Agents in my 🤖 On Robots.txt. Since March, I have updated my Django projects to add more robots.txt rules.

Earlier this week, I ran across this Blockin’ bots. blog post and this example, the mod_rewrite rule blocks AI Agents via their User-Agent strings.

<IfModule mod_rewrite.c>
RewriteEngine on
RewriteBase /
# block “AI” bots
RewriteCond %{HTTP_USER_AGENT} (AdsBot-Google|Amazonbot|anthropic-ai|Applebot|AwarioRssBot|AwarioSmartBot|Bytespider|CCBot|ChatGPT|ChatGPT-User|Claude-Web|ClaudeBot|cohere-ai|DataForSeoBot|Diffbot|FacebookBot|FacebookBot|Google-Extended|GPTBot|ImagesiftBot|magpie-crawler|omgili|Omgilibot|peer39_crawler|PerplexityBot|YouBot) [NC]
RewriteRule ^ – [F]
</IfModule>

Since none of my projects use Apache, and I was short on time, I decided to leave this war to the bots.

Django Middleware

I asked ChatGPT to convert this snippet to a piece of Django Middleware called Super Bot Fight. After all, if we don’t have time to keep up with bots, then we could leverage this technology to help fight against them.

In theory, this snippet passed my eyeball test and was good enough:

# middleware.py

from django.http import HttpResponseForbidden

# List of user agents to block

BLOCKED_USER_AGENTS = [
    "AdsBot-Google",
    "Amazonbot",
    "anthropic-ai",
    "Applebot",
    "AwarioRssBot",
    "AwarioSmartBot",
    "Bytespider",
    "CCBot",
    "ChatGPT",
    "ChatGPT-User",
    "Claude-Web",
    "ClaudeBot",
    "cohere-ai",
    "DataForSeoBot",
    "Diffbot",
    "FacebookBot",
    "Google-Extended",
    "GPTBot",
    "ImagesiftBot",
    "magpie-crawler",
    "omgili",
    "Omgilibot",
    "peer39_crawler",
    "PerplexityBot",
    "YouBot",
]

class BlockBotsMiddleware:

    def __init__(self, get_response):
        self.get_response = get_response

    def __call__(self, request):
        # Check the User-Agent against the blocked list
        user_agent = request.META.get("HTTP_USER_AGENT", "")
        if any(bot in user_agent for bot in BLOCKED_USER_AGENTS):
            return HttpResponseForbidden("Access denied")
        response = self.get_response(request)
        return response

To use this middleware, you would update your Django settings.py to add it to your MIDDLEWARE setting.

# settings.py

MIDDLEWARE = [
    ...
    "middleware.BlockBotsMiddleware",
    ...
]

Tests?

If this middleware works for you and you care about testing, then these tests should also work:

import pytest

from django.http import HttpRequest
from django.test import RequestFactory

from middleware import BlockBotsMiddleware

@pytest.mark.parametrize("user_agent, should_block", [
    ("AdsBot-Google", True),
    ("Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)", False),
    ("ChatGPT-User", True),
    ("Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/58.0.3029.110 Safari/537.3", False),
])
def test_user_agent_blocking(user_agent, should_block):
    # Create a request factory to generate request instances
    factory = RequestFactory()
    request = factory.get('/', HTTP_USER_AGENT=user_agent)

    # Middleware setup
    middleware = BlockBotsMiddleware(get_response=lambda request: HttpResponse())
    response = middleware(request)

    # Check if the response should be blocked or allowed
    if should_block:
        assert response.status_code == 403, f"Request with user agent '{user_agent}' should be blocked."
    else:
        assert response.status_code != 403, f"Request with user agent '{user_agent}' should not be blocked."

Enhancements

To use this code in production, I would normalize the user_agent and BLOCKED_USER_AGENTS variables to be case-insensitive.

I would also consider storing my list of user agents in a Django model or using a project like django-robots instead of a hard-coded Python list.

I spent a lot of time in my shell (terminal, iTerm, bash, whatever you prefer), and by and large, my least favorite and most frustrating chore has always been trying to customize my shell prompt.

A few years ago, I tried to switch from BASH to Zsh and even the Fish Shell, and I quickly reverted because I found customizing my prompt to be such a chore.

Starship is one of those few life-changing development utilities that quickly solved how to customize my shell prompt. Starship also works across various shells, which means changing shells doesn’t mean starting over with a new prompt. Even though Fish and Zsh never stuck for me, Starship was a keeper.

Starship’s default prompt is really good and is a nicer starting point than Zsh or BASH gives you by default. Starship includes a bunch of default settings, which enables support for tools like aws, battery percentage, bun, git, k8s, node js, python, etc, so most of my dotfiles config is turning off the features that I don’t care about.

There is a Dracula Theme for Starship, which is my go-to theme for any application that I can find that supports the color scheme.

Setting up and configuring Starship is a breeze. The documentation is clear and concise, and Starship’s settings are neatly organized in one readable and updated file, ~/.config/starship.toml. Best of all, any changes you make are immediately applied, eliminating the need to restart your shell session. With Starship, you can confidently customize your prompt and see the results instantly.

If you use Homebrew, brew install starship will install it. You can activate Starship by adding eval "$(starship init bash)" to your ~/.bashrc file.

If you have been thinking about trying out Starship or need help figuring out where to start or how to customize your shell prompt, I highly recommend it.

Last year, I wrote this TIL: Creating a Rustywind pre-commit hook, which illustrates how to use the Rustywind CLI, which helps organize Tailwind CSS classes.

Rustywind is a CSS class linting and formatting CLI tool that helps sort Tailwind CSS classes consistently and can help to de-duplicate them. If you struggle with duplicate classes or feel overwhelmed, Rustywind might be what you need.

I prefer to install Rustywind as a pre-commit hook to run on every git commit or when I run pre-commit run.

This .pre-commit-config.yaml config will run the latest version of Rustywind every time pre-commit runs.

# .pre-commit-config.yaml
repos:
  - repo: local
    hooks:
      - id: rustywind
        name: rustywind Tailwind CSS class linter
        language: node
        additional_dependencies:
          - rustywind@latest
        entry: rustywind
        args: [--write]
        types_or: [html]

pre-commit is also an excellent way to manage your project’s other linting and formatting tools. It can be used to ensure you always use the latest version of Rustywind and other tools.

Family

The mowing season has officially kicked off. I mowed our yard for the first time this year on the highest setting because our grass seemed to have grown a foot in a week in several places since I fertilized it last weekend. I also trimmed and removed leaves and branches from the yard. I was rewarded by a late afternoon migraine, but lucked out and only lasted for a few hours.

I finally broke out my miter saw and portable workbench and made room for it in the garage. As I fixed one of the beams that held up the front porch, I realized the water damage was much worse than I thought. It was only held up by one beam, so I’m glad I didn’t put off fixing it any longer. The new support is now fixed, primed, and ready for some wood filler in the holes with a fresh coat of paint to be applied.

My son was an excellent helper this weekend, spending most of Sunday following me around and helping. We also went to a Lawrence fixture Cotton’s Hardware store to get some grass seed, windshield wiper fluid, and a replacement garden hose.

My daughter discovered a new Batman cartoon and spent Sunday shouting, “I am the night” and “I am Batman” as well as any two-year-old can.

Work

It was a busy work week, juggling two client projects. We had a client demo and then a design handoff, in which we now have to turn a static PHP website into a fully working Django website.

Community

Django Code of Conduct WG: I spent more time than expected this week helping with this. We were down to just me, and now we have at least three, soon four, and soon more people to help out.

Django Packages was accepted as a Djangonaut Space project, so I spent time preparing for some June deadlines.

Trailhawks: I met with my running club’s new President, and we discussed some changes to help get everything ready for the new 2024 racing season.

DEFNA: I walked the board through the conference budget, what to look for, and how to know how we are doing.

Conference Chats: We had our monthly Wednesday night meetup.

Office Hours: I hosted a Friday session.

Side projects

Django News Newsletter: We shipped issue #227.

Django News Jobs: I refactored some of my django-fsm states with a few different workflows to automate importing jobs via the aggregator.

Writing

I wrote and published every day.

2024-04-14: 🧱 Dear Tech Companies: Two-year-old edition 

2024-04-13: 🙋 How often do you use Django’s startproject and startapp? 

2024-04-12: 📅 Office Hours on April 19th 

2024-04-11: 📌 On leaving Pinboard, aka Raindrop, is better 🤷 

2024-04-10: 🐛 Production bugs happen to everyone 

2024-04-09: 📋 Copy All Urls browser extension 

2024-04-08: 🐳 Docker and OrbStack disk cleanup notes 

“Me” time

I spent a night deep diving on Simon’s Building files-to-prompt entirely using Claude 3 Opus article, which meant registering a paid Claude account, paying for API access, and spending $0.50 to try everything out.

The results were better than expected, but the AI bar is still pretty low. It helped write tests and some light refactors, but it’s mostly a copy-and-paste-to-refactor tool vs. a write-out working code tool.

Entertainment

📺 Fallout is excellent and worthy of all the praise.

📺 X-Men ‘97 is a nostalgic ’90s throwback cartoon. If I hadn’t watched this series as a teenager, I’m not sure I would have liked it as an adult.

🎮 Double Dragon Gaiden: Rise of the Dragons - Sacred Reunion: My son and I picked up the DLC and started playing this weekend. The game is fun but only has five levels, which is a bit short. We can beat it in one session, which is nice, but the DLC improves on the original with a survival mode, which we liked a lot.

Next week

More work. More house projects. More office hours. More community projects. Figure out my PyCon US plans.

Dear Tech Companies,

There is never a good time to interrupt a two-year-old to ask her for a one to five-star review while she’s playing the LEGO® DUPLO® WORLD game.

I understand that app review ratings are a vital metric impacting your rank in App Stores. While it pains me to write this, if your app is good and my children like it, we parents will rate it on our own time instead of being forced to so our children can return to their firetruck game because an App Store review popup prevents them from playing. I assume no one on the LEGO app store team has children to understand the pain, confusion, and agony this brings.

It’s deeply frustrating to encounter in-game purchases in a game designed for TWO TO FIVE-YEAR-OLDS. Fortunately, we’re on the Apple One / Apple Arcade plan, which allows us to bypass such unnecessary features.

I hope that you can learn from the mistakes made in the LEGO® DUPLO® WORLD game. This way, you can avoid frustrated parents leaving you with low app store ratings in the heat of the moment.

I saw this Mastodon post last week and thought it worth sharing.

Weekend #django question/poll

How often do you use startproject & startapp or something else?

These commands vary by the client or project I’m working on.

I use startproject one to two times a month.

I use startapp one to a dozen times a week.

I have used copier and cookiecutter, and both have their place, but I only use them occasionally.

I create more new projects and apps than the average developer because I’m curious and spin up projects quickly to try out ideas. I prefer this over every experiment being part of a mono project because I will throw most of these away.

and how many years since you started using django?

I have been using Django since 2006.

📅 Next week, I’m hosting Office Hours on Friday, April 19th, 2024, from 2 PM to 4:30 PM Central Time (US and Canada). If you need to convert it to your local time zone, use this handy link.

ℹ️ During our session, we will discuss our ongoing projects, catch up with one another, and work together to wrap up our week on a productive note.

🙏 As always, everyone is welcome to join, whether you’re a regular attendee or joining for the first time.

✅ If you need any additional details, feel free to send me a message or check out the gist from our previous sessions, where you’ll find the Zoom link ⚠️

I look forward to seeing everyone.

A friend asked me which bookmarking service I switched to after I moved on from the Pinboard bookmark service.

tl;dr I use Raindrop. The product and API works wonderfully. 👍

A good bookmark service should:

• be easy to add a bookmark
• have a good browser extension or have a good bookmarklet
• be the best personal search engine
• be a good research tool for taking notes and allowing tagging
• be reliable
• maintain reasonable uptime

Today, I realized I had been on Raindrop’s FREE plan for over a year, so I fixed that and upgraded to a PRO account.

That’s a good sign since I have ~40k and over two decades of bookmarks, and they never pressured me to upgrade or put limits on my account in any frustrating way.

Why I gave up on Pinboard

I could write thousands of words about how frustrating being a paid user was, but I’ll sum it up as running Pinboard wasn’t a priority for Pinboard’s founder, and it painfully showed. 👎

I have noticed a GitHub bug for a few weeks, but I kept forgetting to ask publicly about it.

Today, I noticed it again, and I tracked it down to README.rst files, so I wrote this post:

Has anyone else noticed that some READMEs on GitHub have been broken recently?

I thought it was just me, but the pattern seems to be like code blocks in rst files no longer get formatted.

https://mastodon.social/@webology/112249549385519057

This led to several friends chiming in and one link to this open GitHub issue thanks to Funky Bob

The story’s moral is that writing software is hard; even GitHub makes mistakes and breaks things.

If a trillion-dollar tech company can break things for weeks, cut yourself some slack the next time you stumble on a production.

Disclaimer: I wrote this post while talking in this month’s Conference Chats.

If you are like me and have a habit of opening links in new tabs for days, weeks, and sometimes months, then you probably have not discovered the Copy tabs URL browser extension.

The Copy tabs URL browser extension is a lifesaver for those who habitually open links in new tabs. It effortlessly copies all open tabs from a browser window into your clipboard, eliminating the need for manual copying and pasting. This makes it a breeze to paste into your editor of choice, such as Obsidian, saving you valuable time and effort.

With the Copy tabs URL browser extension, I can easily copy all of my open tabs into a markdown document. This simple action helps me overcome mental hurdles, as I know that the link is safely filed away where I can search and find it later. This makes closing the tab a much easier decision for me.

For longer-term storage, I use the Raindrop bookmark manager to track links I want to find later because it doubles as an excellent research tool and personal search engine. Before Raindrop, I frustratingly used Pinboard for a decade.

So, while the Copy tabs URL browser extension works with Chrome, I switched to the Vivaldi browser last year, and it works just fine.

Docker is one of my daily drivers, and here are my notes for reclaiming disk space and a bonus tip for getting a faster runtime.

Docker “Containers” Prune

Most of the time, when I run into a Docker issue, it’s because I have out of disk space allocated for Docker containers and images.

90% of the time, you can reclaim the most disk space by running docker system prune.

Why? Docker’s default behavior is to remove “stopped containers,” which are instances of containers that have already run and are no longer valid. There is no good reason to keep these around, so I consider this a safe operation to run.

$ docker system prune
WARNING! This will remove:
  - all stopped containers
  - all networks not used by at least one container
  - all dangling images
  - unused build cache

Are you sure you want to continue? [y/N] y
Deleted Containers:
...
Total reclaimed space: 127.5GB

I freed up 127.5 GB of disk space when I ran this today.

Docker “Images” Prune

The docker system df will show us how much disk space Docker uses and what storage types.

$ docker system df
TYPE            TOTAL     ACTIVE    SIZE      RECLAIMABLE
Images          100       0         64.76GB   64.76GB (100%)
Containers      0         0         0B        0B
Local Volumes   1211      0         32.79GB   32.79GB (100%)
Build Cache     638       0         31.29GB   31.29GB

The `docker image prune -a' will delete all images that don’t have a container associated with them.

Why? Worse case, if you delete an image that you need, you can always rebuild it. If you are using a third-party image, you can always docker pull a more recent image. I consider this a safe operation to run.

$ docker image prune -a
WARNING! This will remove all images without at least one container associated to them.
Are you sure you want to continue? [y/N] y
...
Total reclaimed space: 11.28GB

The results after we have run our command:

$ docker system df
TYPE            TOTAL     ACTIVE    SIZE      RECLAIMABLE
Images          0         0         0B        0B
Containers      0         0         0B        0B
Local Volumes   1211      0         32.79GB   32.79GB (100%)
Build Cache     638       0         31.29GB   31.29GB

Docker “Build Cache” Prune

Docker’s Build Cache stores or caches all of the various build steps that Docker will use to assemble an image.

I consider this a safe operation because the worst-case scenario of a build step not being cached is that it takes a little longer to run your next Docker build while Docker re-runs the command.

$ docker builder prune
...
Total:  31.29GB

$ docker system df
TYPE            TOTAL     ACTIVE    SIZE      RECLAIMABLE
Images          0         0         0B        0B
Containers      0         0         0B        0B
Local Volumes   1211      0         32.79GB   32.79GB (100%)
Build Cache     2         0         0B        0B

Docker “Local Volumes” Prune

Docker volumes are where your persistent data lives, like the files your database needs between sessions.

I do not recommend bulk deleting all of your volumes.

Unless you really, really mean it, and you know what you are doing, leave docker volume prune alone.

I have “Local Volumes” that have lived longer than some Javascript Frameworks have.

Instead, only delete what you know you aren’t using anymore via:

$ docker volume ls
$ docker volume rm {volume-name}

OrbStack for extra performance

If you are on macOS, try out OrbStack, which lives up to its claim of being lightning faster than Docker.

I run both Docker and Orbstack on my machines, and I use a DOCKER_CONTEXT environment variable to change between the two contexts.

# for OrbStack
export DOCKER_CONTEXT=orbstack
$ docker system df
...

# for Docker
export DOCKER_CONTEXT=default
$ docker system df
...

If you aren’t sure what your “context” options are, you can run docker context ls to see them:

$ docker context ls
NAME              DESCRIPTION                               DOCKER ENDPOINT                                        ERROR
default           Current DOCKER_HOST based configuration   unix:///var/run/docker.sock
desktop-linux *   Docker Desktop                            unix:///Users/username/.docker/run/docker.sock
orbstack          OrbStack                                  unix:///Users/username/.orbstack/run/docker.sock

Family

I skipped last week’s update because we spent Easter weekend with my family and ~6 hours in the car with the kids.

We went to our favorite local nursery, Vinland Valley Nursery, to pick up some new Hostas and other heartier plants since April is still a bit early for planting flowers because we typically have a few late freezes through Spring. We planted everything, and I fertilized the lawn to prepare it for the season.

I unboxed and assembled my portable workbench and miter saw. My miter saw shipped with everything but the manual, and by the time I found one and printed it, it was time for me to make dinner. I cooked burgers on the outdoor griddle for the first time this year.

Work

I observed Good Friday on Monday and then worked a full day on Saturday to keep up with two client projects.

Community

My community cup was overflowing this week:

• Monthly DEFNA Board of Directors meeting
• Biweekly PSF Conduct WG meeting
• I interviewed with Djangonauts to see if Django Packages was a good fit for a future session.
• I joined one of Wiggle Work’s lunch sessions.
• I joined Conference Chat’s Friday session over lunch.
• I hosted Office Hours for my first meeting back in a month.
• I am working on rebooting the DSF’s Code of Conduct WG and getting it back up to a healthy number of people.
• I attended half a day of PyCascades

Side projects

Mastermind group: A friend invited me to attend their monthly Mastermind business group, which was interesting. We set some goals for next month.

Django News Newsletter: We shipped issue #226.

Django News Jobs: I wrote up a few new features, including potentially moving the project to Django FSM, which might save me a few headaches. I’m also debating a new feature to bring Company Endorsements into the listings.

Django Packages: Security updates. I gave Ryan some reviews on an upcoming profile/extra links feature he’s working on. It’s pretty cool. I drafted a few issues for some potential new feature ideas.

Upgrade Django: I bumped to list the latest few Django releases.

Writing

I wrote and published every day.

2024-04-06: 🚜 Refactoring and fiddling with Django migrations for pending pull requests 🐘 

2024-04-05: ⛳ Syncing Django Waffle feature flags 

2024-04-04: ⬆️ The Upgrade Django project 

2024-04-03: 🚀 Mise is really fast 

2024-04-02: TIL: Brandolini’s law

2024-04-01: 🔥 An update on OpenSSF’s XZ fumble 

“Me” time

Even though March is over, March Madness continues, and I enjoyed two more weekends of Men’s and Women’s basketball. I took first place in my women’s bracket this week. My son tied for #2 and was excited about it tonight.

While I am in first place for my men’s bracket, they play tomorrow (Monday), and the team I picked to win lost last weekend. So I will only win this year if there is a team fight and both teams forfeit. So there is still a chance.

Next week

I should have mowed today and made another pass at removing sticks and leaves accumulated over the Spring. So I suspect we’ll spend some time outside doing that next week.

It’s also time to break out the miter finally saw and spend half a day replacing old boards, repairing fences, and getting my money’s worth out of it.