Annika Backstrom - sixohthree.com

What's in Your Kit? 2026

2026-04-18T00:00:00Z

I started list late last year but never finished it! Anything that's changed since then is struck through with the replacement listed.

I have added a /uses page for this list, which may be more up-to-date than this page.

Hardware

Fairphone 6 running Murena /e/OS (de-Googled Android). Card holder accessory so I can lose all my important stuff at the same time.

~~MacBook Pro (2020)~~~~, Core i7. Feels pretty slow on modern macOS but i don't~~ ~~relish the idea of buying a new Apple computer.~~ Lenovo ThinkPad X1 Carbon Gen. 12 (16 GB RAM, 500 GB HD) with Ubuntu 24.04. My daily driver.

Boox Palma, Android e-ink device in a phone form factor.

MNT Pocket Reform open hardware laptop. I don't use this a ton, despite wanting it to be a part of my regular workflow. Not having a suspend function makes it a poor fit for my distracted lifestyle.

reMarkable 2 e-ink tablet, rarely. My son has commandeered this for drawing.

AMD Ryzen 7 PC with Nvidia 4070 RTX. For gaming and dev work using Windows Subsystem for Linux (WSL).

My work provides me with a MacBook Pro M4. I use a Logitech StreamCam, M-AUDIO M-Track Solo, Marantz Professional MPM-1000 condenser mic, and Sennheiser HD 599 headphones (those open back ones in the ivory and brown colours).

Software

Neovim with LazyVim for coding.

Termius (sparingly) for mobile SSH.

Signal (often) and Delta Chat (rarely, though I wish this was my primary) for encrypted chat.

Synology Photos for photo organisation. I would like to self-host Immich some day.

FreshRSS (self-hosted) for feed reading, self-hosted.

Gitea (self-hosted) for git repositories. Possibly moving to Forgejo on source.tube (part of omg.lol).

Tusky for GoToSocial and Mastodon.

Convos (self-hosted) for IRC.

~~Firefox~~ Zen Browser (Firefox fork) for web browsing, Lagrange for Gemini and Gopher.

Websites

DuckDuckGo for search.

Vanishingly few regular websites these days. A sad state of affairs.

Miscellaneous

Nordace Aerial Infinity Flap backpack, 22L.

Pinned Tabs in Zen Browser

2026-01-03T00:00:00Z

I recently switched from Firefox to Zen, since Mozilla won't stop fucking around with inane AI features and Google cannot be trusted. Zen has some great features, including a UI mostly shoved into the sidebar, pop-up pages, and split tabs. It also handles pinned tabs really nicely but I didn't grok them right away, so I'm writing up my findings here.

In Firefox, I have a bookmark bar with folders of frequently-viewed pages: Jira boards, collections of pull requests, dashboards, etc. Clicking on a bookmark opens an ephemeral tab that I close when I'm done. I also have a few pinned tabs for things like email.

Zen, on the other hand, organises pinned tabs in two different ways, with different visual treatment and ways of handling tab persistence.

At the top we have "Essentials," a grid of pins (the blue dashed box in the above screenshot). They are somewhat similar to pinned tabs in Firefox.

Clicking an on-site link will navigate within the tab.
Clicking an off-site link in the tab will open a pop-up "Glance" page.
Closing or "unloading" the tab will reset the pin to its original location.

The tabs between Essentials and the "New Tab" button (red solid box) are also pinned tabs, but they work a bit differently, especially when organised into folders:

Same behaviour as Essentials with regards to clicking links and resetting the tab.
Collapsing a folder of pins will hide child tabs and sub-folders.
Hovering the folder shows a popup with the folder contents.
Clicking a tab in the popup will show that tab under the collapsed folder, until that tab is unloaded. The various "Switch to Tab" methods that open a pin will act the same way.
Tabs can be unloaded individually or folder-by-folder.
A pinned tab title can be renamed by double-clicking the name in the sidebar.

In aggregate this feels really different from how I use the Bookmark Bar and pinned tabs in Firefox, but I'm a total convert after a day using it. Many of the pages I visit in a workday are recurring and temporary. Zen lets me load my pull requests tab, click around to do my work, and Cmd-W to close the tab to reset it for next time. I can even bind a key to quickly reset the pinned tab to the original URL. This keeps the page confined to a single consistent part of the sidebar, which is a real boon to the way I want to work.

Props to the Zen folks for a really great addition to the browser landscape.

Migrating from Logseq to Trilium

2025-12-16T00:00:00Z

Here's how I migrated my 500-something page knowledge graph/base/whatever from Logseq to Trilium. Just want the script? Download trilium-import.py.

Importing Logseq into Trilium is not streamlined. You can zip up parts of your Logseq directory and import them through the Trilium app, but the notes will retain Logseq's bullet point structure and document links will break. Block-level meta for things like Cards will be inlined into the bullet point. Document links are replaced with the text "missing note." The content comes over but overall it's a mess.

Trilium itself recommends upload_md_folder from trilium-py, but this is only marginally better than the manual import. Linked documents retain their names instead of saying "missing note" but they are not links. This still didn't feel like a good starting point.

logseq2trilium did a good job of formatting my Logseq graph for Trilium. It splits bullet points into paragraphs, which is a better fit for Trilium. The links come across properly. Images linked in documents work after importing. The journal files also have a better name.

Cleaning up the import

Once imported, everything was tucked away inside its own top-level folder. I needed the 3 years of journal files to fit in with Trilium's Journal: year folders, with nested day-of-month folders, with journals inside.

As a first step, I selected all the journal notes in the Trilium app and manually applied the template "Day Note Template" using Bulk Actions > Add Relation: Template.

I wrote a script trilium-import.py that uses trilium-py to clean up the imported journal files:

Create a set of year notes with nested month notes (this does not happen lazily)
Grab all imported child notes from a hardcoded parent note ID
Update titles to match the format "10 - Monday" (day of month, weekday)
Add journal note attributes
Move notes to correct spot in the hierarchy

The import felt much more manageable after getting the journal sorted, and I could get to the more enjoyable part of reorganising and settings icons.

trilium-import.py

from trilium_py.client import ETAPI

import calendar
import pprint

TOKEN_ID = "8EnnBezl2LMP_oa4bkgdp9oG/CYRjyl+MihBIJqbB5HcwipVomH+ovSU="
JOURNAL_ID = "TPz1EGZtugH3"
IMPORTED_JOURNAL_PARENT_ID = "Sy22f3PKxxka"

# Years to create (note that stop value is not included)
YEARS = range(2022, 2026)

def has_attr(note, attribute):
    return any(attr["name"] == attribute for attr in note["attributes"])

class Graph:
    def __init__(self):
        self.notes = {}

        server_url = 'http://localhost:37840'
        self.ea = ETAPI(server_url, TOKEN_ID)

    """Find a note by its title."""
    def find(self, parent, name):
        res = self.ea.search_note(
            search="note.title = '%s'" % (name),
            ancestorNoteId=parent,
            fastSearch=False,
            orderBy=["title"],
            limit=1,
        )

        if res['results']:
            return res['results'][0]

    """Find single note by its ID, or raise an exception."""
    def find_by_id(self, id):
        res = self.ea.search_note(
            search="note.noteId = '%s'" % (id),
            fastSearch=False,
        )

        if res['results']:
            return res['results'][0]

        raise RuntimeException("could not find note with id %s" % (id))

    """Find a note by its name or create it."""
    def find_or_create(self, parentNoteId, title):
        note = self.find(parentNoteId, title)

        if not note:
            note = self.create_empty_note(parentNoteId, title)

        return note

    """Create an empty note."""
    def create_empty_note(self, parent, name):
        res = self.ea.create_note(
            parentNoteId=parent,
            title=str(name),
            type="text",
            content="..."
        )

        # For some reason, we can't use an empty content above.
        self.ea.update_note_content(res["note"]["noteId"], "")

        return res["note"]

    """Set a label attribute if the attribute does not already exist."""
    def add_label_if_missing(self, note, name, value):
        if not has_attr(note, name):
            self.ea.create_attribute(
                noteId=note["noteId"],
                type='label',
                name=name,
                value=value,
                isInheritable=False
            )

    """Create the requested year folders, if they do not exist.

    Set missing attributes for yearNote and sorting.
    """
    def populate_years(self):
        # It would be better to create this lazily based on incoming notes.
        for year in YEARS:
            yearNote = self.find_or_create(JOURNAL_ID, year)

            self.add_label_if_missing(yearNote, "yearNote", year)
            self.add_label_if_missing(yearNote, "sorted", "")

            self.notes[year] = (yearNote, {})

            self.populate_months(year)

    """Create month folders for a year, if they do not exist.

    Set missing attributes for monthNote and sorting.
    """
    def populate_months(self, year):
        (yearNote, months) = self.notes[year]

        for i in range(1, 13):
            title = "%02d - %s" % (i, calendar.month_name[i])

            monthNote = self.find_or_create(yearNote["noteId"], title)

            self.add_label_if_missing(monthNote, "monthNote", "%d-%02d" % (year, i))
            self.add_label_if_missing(monthNote, "sorted", "")

            months[i] = monthNote

    """Generator to return all children of a given note."""
    def imported_journals(self, parentId):
        results = self.ea.search_note(
            search="note.noteId = '%s'" % (parentId),
            fastSearch=False,
            orderBy=["title"],
        )

        if len(results['results']) == 0:
            raise RuntimeException("cannot find parent journal by ID")

        parentJournal = results['results'][0]

        for childId in parentJournal['childNoteIds']:
            yield self.find_by_id(childId)

    """Given a journal note, fix titles and add journal attributes, and
    relocate to Journal tree."""
    def relocate(self, note):
        title = note["title"]

        (date, weekday) = title.split(" ")
        (year, month, day) = date.split("-")

        year = int(year)
        month = int(month)
        day = int(day)

        newTitle = "%02d - %s" % (day, weekday)

        print("{0} -> {1}".format(title, newTitle))

        self.add_label_if_missing(note, 'dateNote', '%04d-%02d-%02d' % (year, month, day))

        self.ea.patch_note(noteId=note["noteId"], title=newTitle)

        monthNote = self.notes[year][1][month]

        self.ea.create_branch(
            noteId=note["noteId"],
            parentNoteId=monthNote["noteId"]
        )

        self.ea.delete_branch(
            branchId="%s_%s" % (IMPORTED_JOURNAL_PARENT_ID, note["noteId"])
        )

if __name__ == "__main__":
    graph = Graph()

    print("Populating years... ", end='')
    graph.populate_years()
    print("done.")

    for note in graph.imported_journals(IMPORTED_JOURNAL_PARENT_ID):
        graph.relocate(note)

Fediverse Rules of Engagement

2025-10-18T00:00:00Z

These are my "rules" for how I interact on the Fediverse (Mastodon, GoToSocial, and other ActivityPub friends). I don't claim that they're perfect or correct or appropriate for anyone other than me.

Follow and unfollow liberally. Don't be afraid to follow someone new, even if I'm happy with my feed. Follow folks boosted into my feed, follow folks who interact with me, follow from the federated timeline. Unfollow if I want to, but consider using private notes, hiding boosts, and filters as "strikes."

Check remote replies before sending my own reply. Every server has a partial window of the Fediverse. Is my reply redundant? Is the person already drowning in replies?

Do not meta. Meta is the mind killer. Meta is the little death that brings total obliteration.

Don't overload posts with disclaimers. They clutter up the post and deprive me of the opportunity to block jerks.

Always post alt text. Do not boost anything without alt text, but don't harass individuals for not writing alt text.

Treat unverified accounts of "famous" people as illegitimate. Misinfo is too rampant to risk engaging with impersonators.

Don't boost news that lacks a clear citation. Be wary of screenshots of text that appear to be from legit sources, or are from unidentifiable sources.

No GenAI. Don't boost, don't interact, consider muting or blocking the poster depending on severity. Be extremely wary of "activists" posting GenAI slop.

Eleventy Migration Notes

2025-10-13T00:00:00Z

In lieu of a retrospective of my weekend-long mad dash to convert this site from Pelican to Eleventy (11ty), here are a few pages I found helpful and some things I learned along the way.

Neovim supports a .lazy.lua file for project- or directory-specific settings.

Working with custom taxonomies in Eleventy means jumping through a couple hoops but it's not impossible. Once you get a handle of the Collections API, it's alright.

It's helpful to have a script for manipulating front matter. Here's mine.

Links (16 July 2025)

2025-07-16T00:00:00Z

DIY Book Lamp
PiKVM
28 slightly rude notes on writing
JSON Structure -- JSON Structure is a data structure definition language that enforces strict typing, modularity, and determinism.
Fedify -- A TypeScript library for building federated server apps powered by ActivityPub and other standards, so-called fediverse
The ‘Sinners’ Movie Syllabus via @jalcine@todon.eu
Slidev -- Presentation slides for developers
Charm -- Various commands and helpers for pretty interactions in your CLI
Conventional Commits -- A specification for adding human and machine readable meaning to commit messages
Flattening Rust's Learning Curve
PechaKucha -- "a storytelling format in which a presenter shows 20 slides for 20 seconds per slide"
SaySomethingin language learning
24hourtime.info
- sunclock.net
Concurrently -- Run multiple commands concurrently
Rio -- A modern terminal for the 21st century
IBM Equal Access Toolkit -- accessibility tools from IBM

Links (22 April 2025)

2025-04-22T00:00:00Z

bylandandsea.ie – Travel to and from Ireland without flying
Garage — An open-source distributed object storage service tailored for self-hosting, supporting the S3 API
Human chains feel like a metaphor in these dark times
- ICE came for their neighbor, so these Tennesseans formed a human chain to protect him
- ‘Book brigade’: US town forms human chain to move 9,100 books one-by-one
stty -ixon to stop ^S from freezing your terminal via @b0rk@social.jvns.ca
Rust for the Polyglot Programmer via @raptor@infosec.exchange
How to Make Chai
The Movie Mistake Mystery from "Revenge of the Sith"
Typst – A more productive workflow for science.
Twine Cookbook – "This Cookbook contains documentation, tips, and examples for using the non-linear story creation tool Twine"

Webhooks as a (Systemd) Service

2025-04-12T00:00:00Z

The "docs" microsite for xoxo.zone is a static page built with Eleventy. I explicitly didn't want to overcomplicate the site's setup with a cloud build process triggered by commit actions: the static site is compiled locally and committed alongside the content changes. A cron on the server runs git pull on the repo every 5 minutes. The web server can directly serve the site without any additional build.

The site is updated a couple times a month on average. Of the 8,640 average monthly cron git pulls, 8,638 will do nothing. The net impact of this is probably negligible, but it did annoy me. Besides, Codeberg (my forge of choice for xoxo) gets the occasional DDoS and I'm sure is getting hammered by "AI" scrapers that never sleep. Why send them more traffic than necessary?

I wanted to update the site when I pushed to the repo, without a complex configuration that would be difficult for someone else to pick up. Maybe I could write a lightweight web server that listened for a request and perform an action?

Stop giving things generic names

The obvious choice for triggering an action on push is a Webhook, which Codeberg supports natively. Expose an endpoint, verify incoming requests, run a command.

Doing some preliminary searches, I stumbled across the extremely generically-named webhook, a Go web server for triggering commands based on webhook requests. It's apt-installable on Ubuntu, has a simple configuration file syntax in JSON or YAML, and doesn't mind being proxied behind nginx. From a maintenance perspective, that's better than rolling my own server.

Webhook configuration

After apt install webhook, I customised some of the launch parameters with systemctl edit webhook^[1]

[Unit]
ConditionPathExists=
ConditionPathExists=/etc/webhook.yaml

[Service]
ExecStart=
ExecStart=/usr/bin/webhook -nopanic -hooks /etc/webhook.yaml -ip 127.0.0.1 -port 9899 -urlprefix my-webhook-prefix
User=www-data
Group=www-data

I configured my hook in /etc/webhook.yaml. webhook has an awkward syntax for passing arguments to commands, so I made a small bin script to wrap cd and git pull .... webhook also includes "matchers" to further customise the hook based on incoming parameters. This config performs request signature verification based on a shared secret, and filters for push events.

- id: deploy-docs
  execute-command: /usr/local/bin/xoxo-docs-pull
  response-message: ok
  trigger-rule:
    and:
      - match:
          type: payload-hmac-sha256
          secret: my_secret
          parameter:
            source: header
            name: X-Forgejo-Signature
      - match:
          type: value
          value: push
          parameter:
            source: header
            name: X-Forgejo-Event

Then I proxied the requests through Nginx:

upstream webhooks {
    server 127.0.0.1:9899 fail_timeout=5;
}

server {
    # the rest of the server config...

    location ~ ^/my-webhook-prefix/ {
       proxy_pass http://webhooks;
    }
}

That's enough to get a working webhook. No more cron needed!

The future

Mastodon itself supports webhooks, and I'd love to improve our admin hooks in the future. Today we get a Discord message when a report is created, but it's not very readable and there's no ability to update the initial message when the report is actioned. webhook feels like a good starting place to improve that experience.

Technically these are in an Ansible playbook, but I'm simplifying so the code examples are more self-contained. ↩︎

Migrating xoxo.zone to OVHcloud

2025-04-10T00:00:00Z

Mastodon hosted on xoxo.zone is now living on its new server at OVHcloud. We were hosted at Hetzner for just about 2 years, but draconian terms of service and some scary experiences for other communities made the move inevitable. I evaluated a few EU-based hosts, including Netcup and Scaleway. OVH hit the sweet spot of pricing, specs, and terms of use.

We built this shitty

Due to a miscalculation, the server I provisioned on Hetzner used spinning rust HDDs instead of SSDs. This gave us a ton of storage overhead we didn't need, and was also slow as fuck and made many simple things very painful. Even git operations and restarting services could take minutes instead of seconds.

The upgrade to Mastodon v4.2.0 last October was particularly painful. I first upgraded the server from Ubuntu 20.04 to Ubuntu 22.04, and planned to keep going to Ubuntu 24.04. This required a PostgreSQL upgrade from v15 to v17. I started this upgrade at 15:00, and gave up for the day when the database finally finished rebuilding at 01:30. The server was down the whole time. Bummer.

A lot of effort went into making this look effortless

A challenge of running a server like this is needing to know a little bit about everything. It was clear to me that I could do better than 10.5 hours of downtime, but I wasn't sure how to get there. I've done a lot of reading about PostgreSQL migration strategies since October.

The server database is backed up twice a day. The pg_dump takes about 2 hours, and the upload is another half hour, to say nothing of restoring the db on a new host. Not awesome.

A test rsync --checksum of the database took about 80 minutes, even for subsequent rsyncs that (in theory) had to transfer less data.

Replication was daunting, but I stuck with it. In the end it was pretty painless and worked extremely well.

I created a replication role on the old server, xoxo-4:

CREATE ROLE xoxo5 WITH REPLICATION PASSWORD 'secret_password' LOGIN;

And I updated the access rules in /etc/postgresql/17/main/pg_hba.conf:

# Allow replication from xoxo5@10.0.0.5
host    replication     xoxo5           10.0.0.5/32       scram-sha-256

On the new host, xoxo-5, I emptied out the /var/lib/postgresql/17/main directory and enabled replication:

sudo -u postgres pg_basebackup -h 10.0.0.4 -p 5432 -U xoxo5 -D /var/lib/postgresql/17/main/ -Fp -Xs -R

This took a few hours but it was worth every second. Once the backup was done, the results were extremely promising: select * from pg_stat_replication and pgmetrics showed delays in the milliseconds, and spot checking counts in the statuses and accounts tables looked good (other than count(*) taking over 10 minutes on xoxo-4).

In theory the hardest, slowest part was done: the 88GB database was ready for cutover whenever we were.

Not zero-downtime but I remain chuffed

I was emboldened by successful replication and from listening to Eurovision playlists at high volume for the previous hour. After some encouraging words like "why not" and "if you fuck this up maybe i can focus on work," I finalised a migration plan and kicked things off.

I had done a lot of work already at this point:

I had run the server setup guide and finished an initial rsync on some key directories, including the nginx config
A bunch of server config, including Mastodon service files, backup configuration, and crons are in an Ansible playbook, which I had already run
The domain TTL was already ramped down to 60 seconds

Winding down xoxo-4 looked like this:

mastodon-bounce stop all     # wrapper script that runs systemctl on all mastodon services
mastodon-bounce disable all
systemctl disable --now redis-server.service

Bringing things back up on xoxo-5 looked like this:

/root/sync.sh # /home/mastodon/live, /var/lib/redis, /etc/letsencrypt, /etc/nginx
pg_ctlcluster 17 main promote
systemctl enable --now redis-server.service
mastodon-bounce start all
mastodon-bounce enable all
RAILS_ENV=production ./bin/tootctl feeds build
RAILS_ENV=production ./bin/tootctl search deploy

I would also run a few SQL commands to check data consistency:

sudo -u postgres psql -c '\x' -c 'select * from pg_stat_replication'
time sudo -u mastodon psql mastodon_production -c "select count(1) from statuses"
time sudo -u mastodon psql mastodon_production -c "select id, created_at from statuses order by created_at desc limit 10"
time sudo -u mastodon psql mastodon_production -c "select count(1) from accounts"

Job's done

The migration took about 15 minutes from start to finish. Next time, with SSD hosts on both sides, I could probably get it down to seconds or maybe even zero downtime.

I did hit one snag that was obfuscated by caching in the Mastodon service worker: the mastodon account's home directory was created with permissions 0750, and Nginx could not read files in the web directory, causing a lot of busted client pages for about 30 minutes after I "finished" the migration. There's always something.

But still! It's done and I'm happy with how it went.

Let's not do it again for a long time.

XOXO 2024

2024-08-27T00:00:00Z

Badge lineup: 2015, 2016, 2018, 2019, 2024

The final XOXO Festival wrapped up this weekend in Portland, Oregon. It was my fifth XOXO, and it's been 5 years since the previous festival, XOXO 2019.

Five years is a long time. Maybe long enough that, like me, you see a change in yourself. Since 2019, I'm a little frayed around the edges, more cautious, more reserved in the face of an unfriendly world.

I have allowed myself to harden in the past five years. I saw it happening and I fucking leaned in. It was a learned response to a hostile environment. The timeline of this blog is a reflection of that change: withdrawal, retreat, a reluctance to engage.

XOXO is an event, yes. It's a particular point in time with a beginning and an end. But XOXO is also a feeling of curiosity, a sense of wonder, a rejection of cynicism, a community. A reminder. A dream.

XOXO is made of people. This weekend, those people reminded me what it's like to allow myself to feel joy and hope. I saw a lot of old friends and maybe made some new ones. I heard their stories and felt their excitement. Hibernating parts of me woke up, did a little stretch, and thought maybe it's time to leave the cave.

It's easy to harden up. Only natural, as I heard through a friend. Going soft again is not so easy. All these great people make me want to at least try.

There will not be another XOXO Festival, but there are many people who embody its spirit, whether they associate it with the festival or not. I will try to remember that without the festival's periodic reminder.

The traditional "leaving XOXO" shot, 2024

Collected XOXO 2024 Posts

The Final XOXO by Kit Jones (SpideyJ)
some thoughts (but not enough thoughts), post-XOXO 2024 by Georgie
This Will All End in Tears by Zoe Landon
My experience of XOXO 2024 by Matthew Lyon
The Last XOXO: 2024 by Brook Shelley
Thanks, XOXO by Jason Kottke
XOXO 2024 by Phil Dini
My first year with XOXO by Brian Moon
The internet doesn’t have to feel like this by Justin Pot
Lowering Expectations, One Project at a Time by BrianEnigma
Creating Community at XOXO by Jonathan Liu
XOXO taught me it was OK to be weird by Josh Rubenoff
XOXO 2024 by Benjamin Chait
XOXO by Emory Dunn
Spotted: A deep dive into the digital jungle, where the survival of the fittest isn’t about brawn but brains—and a little (a lot of) kindness, too by Višnja Milidragović
It has been a time by Jim Withington
After xoxo by Ním Daghlian
goodbye, xoxo by clarity flowers
Lower your expectations by Johnathan and Melissa Nightingale
Farewell, XOXO by Arun Venkatesan
Photo album: XOXO Festival 2024 by Ian Linkletter