Work on Janusworx

Moved From Papermod to Blowfish

feedback@janusworx.com (Mario Jason Braganza) — Fri, 24 Jul 2026 08:30:00 +0530

Intended Audience

Me.
This is a sorta-kinda living document, a record of my idiosyncratic decisions to first move to, and then bend the Blowfish theme to my needs.

Update (2026-07-24) : RSS updates, Miscellanea, some editing.

Blowfish theme logo

I changed my blog theme from Papermod to Blowfish.

Why move to Blowfish?
#

Papermod, to my mind, is perfection. And I outgrew it. That’s all there is to the why.
Papermod caters to the needs of a blog, and I think my blog is slowly turning into a multi section website.
I want menus, and possibly submenus and Papermod does not do them.
The finished outcome, while not as pretty and uniform as the old Papermod site, makes writing and managing it all, much simpler.

Blowfish specific things for future Jason to note.
#

Big Picture things
#

The theme does a lot. Most of which I don’t think I need.¹
The aim is to match what my site looked with Papermod, as closely as possible.
Everything CSS is in assets/css/custom.css, rather than multiple CSS files as it was with Papermod.
Pagefind powers search on the site. The inbuilt search is disabled in main.toml.
Footnotes use Littlefoot; so they appear as popups, and the need to scroll up and down the page is reduced.

Configuration
#

Blowfish separates everything out in its config. Lots of toml files in there.

Base config is in hugo.toml, theme parameters are in params.toml, menus in menus.XX.toml and language + translation related stuff is in languages.XX.toml ²
main.toml sets the theme, the baseURL and I’ve added a section to limit RSS entries to only the last 50 items.
The author parameters along with the social link set, as well as the logo, are in languages.XX.toml. I have duplicated them. One for EN and one for FR. The logo that it requires, is in the assets/logo directory.
Menus and submenus in menus.XX.toml. Duplicated. The English menus have English names and the French, French ones.
Parameters; the header is set to basic. Every other option creates a sticky header of some sort. Lots of other changes. Most are self explanatory. Turned off “Hero” options everywhere, to stick to a plain, basic layout as much as possible.
My split of the various configuration files for environments (base, development & production) have stayed the same with no changes.

Content Structure
#

Directories and subdirectories for each section.
Make sure each section has a corresponding index.XX.md for static pages or _index.XX.md for list pages. This, I think (but am not sure) has fixed my old, nagging issue of page navigation overflowing. I would keep going next, in the Talks section for example, and when I reached the end, it would show me another entry from some other section. Now the front/back buttons stay restricted to the section they are in.

Archetypes
#

All the archetypes have been slightly tweaked. They now also include a template line for an image. It uses the theme’s figure shortcode for images.

i18n (Internationalisation)
#

The internationalisation options are easier here. Just defining a term somewhere and then referring to it here just works.
For e.g., I use search_placeholder as a term in the search layout and then in i18n/en.yaml I have, search_placeholder: "Search for something …" while i18n/fr.yaml gives the same term a different spin, search_placeholder: "Rechercher des articles …"

Layout Overrides
#

RSS
#

Have tweaked the RSS layout a teensy bit. My version is in layouts/_default/rss.xml, which includes the full text of the post, instead of just a summary.

<!-- <description>{{/* .Summary | transform.XMLEscape | safeHTML */}}</description> -->
 <description>{{ .Content | transform.XMLEscape | safeHTML }}</description>

Have restricted RSS feed generation for posts in the Notes and Miscellanea section. They won’t appear in the RSS feed unless I want them to. Most of them are just streams of consciousness, to help me get on with my day. Only a few are exceptions, that I want published.
I even disable sitemap generation for those “important to me, but useless for thee” posts with this bit of metadata in each post³

sitemap:
 disable: true

So I have then taken advantage of this fact, to exclude these posts from having entries in the RSS feed, by telling Hugo to only look at posts where such a key does not exist; which basically, is everything else.

<!-- {{/* range $pages */}} -->
 {{ range where $pages "Sitemap.Disable" "ne" true }}

Search
#

This pulls in the script and the styles for Pagefind, which I use for searching the blog. Layout is in layouts/_default/search.html.
Also note, then when working on search locally, first create an index with pagefind and move it to the static folder because our normal way of working is to render it all in memory.
- do a hugo build
- then do pagefind_extended --site public
- then mv public/pagefind static/pagefind
- then rm -rf public
- and then start the hugo server like we normally do

Littlefoot
#

Footnotes are powered by Littlefoot. Have used layouts/partials/extend-footer.html to pull in the Littlefoot script and layouts/partials/extend-footer.html to grab the CSS.

Redirect
#

An old redirect template to redirect from one section to the other is still in place in the layouts/redirect directory

Shortcodes
#

Most of the old shortcodes made it over. Retired my image and caption shortcodes in favour of using Blowfish’s version. Also created an hsgallery shortcode, so that I could continue using my old Hugo Shortcode Gallery to create image galleries. The gallery shortcode conflicts with Blowfish’s own, which I don’t like and so won’t use.

Static + Tools
#

Fonts and Littlefoot reside in the static directory, along with all my images.
There’s also a tools directory which contains the Pagefind binary. This does not get published obviously. But the Forgejo action uses this to generate an index for the site and then sends the whole thing off to the VM.

Themes
#

Contains Blowfish and Hugo Shortcode Gallery, added as git submodules. I dropped Hugo Admonitions, because Blowfish has them already.

CSS
#

All in assets/css/custom.css

Import fonts: Valkyrie for the site and Source Code Pro for code.
Tweaked the Janusworx up left on my site, and changed all the nav bar items to use Valkyrie Caps.
Dark mode links were not contrasty enough. So changed the colour and added an underline on hover.
Added a class to the figure tag, so I could center the pics I use.
Added an indent to the numbered list items on the Annual Book List pages, so that the numbers stay under each month.
Style littlefoot footnotes, so that the button is slightly smaller and higher, and the notes use Valkyrie, along with a couple of other tiny tweaks.
Style the Blowfish box, and the results, on the search page
Main thing to note. Use !important; liberally. Changes don’t always cascade down correctly.

Miscellanea
#

Go Templates:
#

Use ```go-html-template as the language hint to the code fence to display Go template code correctly. (like I needed to in this post for e.g.)
If it is pure Go Template stuff with little to no html, another alternative is ```go-text-template
I could not just comment out Go template code with HTML comment markers . The Hugo live server kept crashing on me with weird EOF errors, until I figured this out. The Go template block needs a comment marker inside the the {{ … }} section like so, {{/* … */}}.
E.g. {{/* range $pages */}}

Feedback on this post?
Mail me at feedback at this domain.
P.S. Subscribe to my mailing list!

famous last words! ↩︎
XX is the language code. I have EN for English and FR for French. So menus.fr.toml for e.g. ↩︎
I have modified the notes archetype, so this is always pre filled and set. ↩︎

You Can Now Search this Blog

feedback@janusworx.com (Mario Jason Braganza) — Tue, 23 Jun 2026 14:24:24 +0530

Intended Audience

Me. In case I want to retrace the big idea tomorrow

tldr: https://janusworx.com/search
#

Being the lazyass that I am, I used to depend on search engine indexes to find something on my own blog. (First google, then DuckDuckGo.)
Kushal, having https://search.kushaldas.in/ has always made me jealous though.
And now, the days of finding something relevant on my blog, on the first search are long gone. Search engines today are a pale shadow of what they used to be. The current load of crap actually makes me long for the days of Altavista!

So, I decided to get off my butt, and go do something about it, because everytime I want something, my blog is one of the first places I search.
For e.g. when writing this post, I needed to look up how to add line numbers to my code blocks. Search helped me find it :)

I went to the Hugo website and decided to implement the first search tool¹ they suggested: Pagefind.
Well, I did that since Pagefind took care of the only thing I’m anal about. It keeps everything local and under my control. If not, I’d have kept looking for something else.
What it (Pagefind) does, is parse the site, Hugo builds for me, and create a search index along with some more goodies (an api for search as well as UI components) in a bundle. The rendered html and the bundle together are now the artifact that my site is made up of.

The Plan
#

Get Pagefind and install it into a tools subdirectory in my blog’s directory, so I always have it handy (and the Forgejo action can get its grubby paws on it when it wants to run the publishing workflow)
Create an entry in my menu, so I can go clicky to a search page.
Create said search page.
Make search work.
Make the page resemble the rest of the blog as much as possible.

Installing it and wiring it all up
#

1. Installation
#

This is the easy part. Pagefind is a single binary, which I shoved into the tools subdirectory.

2. Creating an entry for search in the navigation menu
#

I edited my Hugo config (mine is called hugo.yaml), and added search as a menu item.
More shifted downwards, as did its weight.

menus:
 main: # elided config
 - identifier: search
 name: search
 url: /search
 weight: 60
 - identifier: more
 name: more …
 url: /more/
 weight: 70

3. Creating the search page.
#

This was a two step process.

a. I needed a placeholder page that I could render some layout (a template) into.
#

This was search.md which lives in my content directory. This is all it contains …

1
2
3
4
5


---
title: Search
layout: search
url: /search/
---

See the key that says layout on line 3? This is looking for a layout called search which is what we now create …

b. Creating the search layout.
#

Papermod already ships with a layout called search. Which I did not want, because it is meant for use with another local search engine called lunr.js.
Hugo can override whatever file your theme provides, by letting you create your own versions of said file in the layouts directory. Hugo will then look at that instead of the theme’s file.
Over the years I’ve gotten dangerous enough by using Hugo’s lovely channel for discourse. Between that, and Hugo’s documentation as well as Pagefind’s own help page, I rustled up a layout file called search.html which lives in my layouts/_default directory. This is what it contains …

{{ define "main" }}

<link rel="stylesheet" href="/pagefind/pagefind-component-ui.css">
<script type="module" src="/pagefind/pagefind-component-ui.js"></script>


<div class="main">
 <article class="post-single">
 <header class="post-header">
 <h1 class="post-title">{{ .Title }}</h1>
 </header>

 <div class="post-content search-content">
 <pagefind-searchbox id="searchbox"
			 autofocus="true"
			 shortcut="/"
			 placeholder='{{ i18n "search_placeholder" }}'>
 </pagefind-searchbox>
 </div>

 </article>
</div>

<script>
 // we’re waiting for the pagefind-searchbox component to be registered 
 customElements.whenDefined('pagefind-searchbox').then(() => {
 const box = document.querySelector('pagefind-searchbox');
 // shadowRoot input, or a delegatesFocus component may expose it differently
 const tryFocus = () => {
 const input =
 box.shadowRoot?.querySelector('input') ??
 box.querySelector('input');
	// search for input boxes in two different ways above
	// and if it finds one focus it below
 if (input) {
 input.focus();
 } else {
 // component rendered but input not yet in DOM — wait one more frame
	 // wait for the shado
 requestAnimationFrame(tryFocus);
 }
 };
 // try one more time
 requestAnimationFrame(tryFocus);
 });
</script>

{{ end }}

This creates the box on the search page, ready to take input and show results.

Update (2026-06-24): The eagle eyed amongst you, have figured out this code has changed between yesterday and today. I used Pagefind’s documentation to simplify what I had yesterday even more and then got one of the llms to write a script that focuses the input box when the page loads, so that we can just start typing in whatever we need to search.

4. Make the page resemble the rest of the blog as much as possible.
#

Doing things a bit out of order and finishing up with the page first.
Leaned on Pagefind’s docs as well as Chatgpt, Claude and Gemini for this part to create a CSS file, that looks pretty native to my blog. I’ve called it include-pagefind.css and it lives in the assets/css/extended subdirectory.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67


:root {
	/* you don’t really need to use `pf-font`, 
	if you don’t have a custom font.
	for some reason, it wouldn’t render the font I specified here, 
	so I forced it with `!important` */
 --pf-font: your-font-name-here !important; 

 --pf-text: var(--primary);
 --pf-text-secondary: var(--secondary);
 --pf-text-muted: var(--secondary);

 --pf-background: var(--theme);
 --pf-border: var(--border);
 --pf-border-focus: var(--primary);

 --pf-hover: var(--tertiary);

 --pf-mark: var(--primary);

 --pf-border-radius: 8px;

 --pf-input-height: 48px;
 --pf-input-font-size: 1rem;

 --pf-result-title-font-size: 1rem;
 --pf-result-excerpt-font-size: 0.90rem;

 --pf-shadow-sm: none;
 --pf-shadow-md: none;
 --pf-shadow-lg: none;
}


.search-content {
 margin-top: 2rem;
}

pagefind-input {
 display: block;
 margin-bottom: 1.5rem;
}

pagefind-summary {
 display: block;
 margin-bottom: 1rem;
}

pagefind-results {
 display: block;
 margin-top: 1rem;
}


pagefind-results a {
 text-decoration: none;
}

pagefind-results article {
 border: 1px solid var(--border);
 border-radius: 8px;
 padding: 1rem;
 background: var(--entry);
}

pagefind-results article:hover {
 background: var(--tertiary);
}

5. Make Search Work.
#

This is a 3 step process

a. Build the blog.
#

Easy peasy. hugo build and we have our blog rendered in the public subdirectory.

b. Get Pagefind to create its index in the build directory.
#

This is easy as well.

./tools/pagefind --site public

and Pagefind creates a pagefind subdirectory in our public directory where it creates its search index after reading all the html.
All I need to do now, is just …

c. Publish!
#

Ok, I lied a teensy bit. I don’t have to do steps a. and b. because they are part of step c.
The Forgejo action checks out my committed files, pulls in Hugo from a private package repo, builds the site, then runs Pagefind to generate the index² and finally uses rysnc to sync it with the VM. Here’s the relevant section of the action.

# lots of stuff elided …

jobs:
 build-and-deploy:
 runs-on: my-big-box

 env:
 HUGO_VERSION: ${{ vars.HUGO_VERSION }}
 steps: # lots of snipped stuff …
 - name: Checkout the website repo
 uses: actions/checkout@v4
 with:
 submodules: recursive
 fetch-depth: 1
 
 - name: Build website with Hugo
 env:
 HUGO_ENVIRONMENT: production
 run: |
 hugo \
 --gc \
 --minify

 - name: Build the Pagefind search index
 run: |
 tools/pagefind \
 --site public

I thought the indexing would take a lot of time and was hesitant to use it as part of a regularly run Forgejo action workflow, but several runs show that the index generation only increases the time by at a couple of seconds at most. This is amazing!

Now I can search my own words, on my own blog, to my hearts content!

Feedback on this post?
Mail me at feedback at this domain.

P.S. Subscribe to my mailing list!

as on 2026-06-23 ↩︎
this is why i chucked it into the tools subdirectory and committed it. ↩︎

Matthias Endler Built His Own Newsletter Setup

feedback@janusworx.com (Mario Jason Braganza) — Fri, 22 May 2026 16:13:29 +0530

Intended Audience

Mostly me (confirmation bias. slight gloating 😂)

I’m writing this post today, because I monkeyed with my newsletter template. Again!
So this post is just to make sure there are no more hiccoughs :)

Matthias Endler, he of the fantastic Rust in Production podcast (and the Corrode rust consultancy) just published a post on his newsletter setup and how he went about building it.

His reasons (and his story) are really similar to mine.¹
Tinyletter => Mailchimp =>
Wanting to have a semblance of control after being burnt => Self hosting!

He settled on Plunk, while I use my Sendy + Amazon SES combo. (they look pretty similar to me)
Newsletters live in a git controlled repo. Same.
Write in my favourite editor. Ditto.
One markdown file per letter. Same pinch!

The only difference, is his workflow is controlled by a fancy cli, while I depend on ye old copy & paste.

Having someone with a visible presence do this, hopefully serves as an inspiration to others.
I just wish more of my nerd kinfolk do this!

Feedback on this post?
Mail me at feedback at this domain.

P.S. Subscribe to my mailing list!

except for his audience ofcourse! Which is orders of magnitude larger than my dinky little circle of friends ↩︎

Found MJML

feedback@janusworx.com (Mario Jason Braganza) — Thu, 21 May 2026 10:51:49 +0530

Intended Audience

Mostly me.
Also hopefully, other folks struggling with, or looking for a way to create email newsletters.

I recently upgraded Sendy, the software, I’m using to create and send my newsletters.
(I don’t need the new shiny, or the extra features, but also those folks have been pretty good as far as pricing and functionality go, and I want them to stick around for a long time.)

The new release, in addition to a refreshed interface, now features a block WYSIWYG editor. And my old template (based on the venerable Antwort) that I used to compose my emails broke. What used to be a predictable way of writing emails, now turned into something filled with papercuts. Everything would look alright, but everything was ever so subtly out of place.

So I went looking for something that I could live with long term once again.
Since all of this is a personal hobby and I only use all of this to send mail to a handful of my friends, most of the paid options go right out of the window, because. And because I am a control freak, hosted options aren’t a thing for me either.

Enter MJML.
From their website:

MJML was created in early 2015 by a team of Mailjet developers while they were working on Passport, Mailjet’s drag-and-drop interface for creating responsive emails. Having been knee deep in email for five years, the Mailjet team saw two things: a) email HTML is antiquated and not developer friendly b) a growing trend of email being viewed on mobile and tablet and the number of screens is only increasing. This means that finding a way to code responsive email easily and quickly is pretty important. The team started by creating a new markup language that would abstract the complexity of responsive HTML and automatically generate it. And that’s how MJML was born.

After learning so much from our users, we wanted to give back to the community by sharing this open-source framework to make responsive email easier and redefine the coding experience once and for all.

It works in the same vein as revealjs.¹
You install the tool and use their semantic markup to write a document.
The MJML cli tool, in this case, then renders it to email worthy HTML.
The markup is simple and well thought out. Writing it is a breeze.
I used Timo Reymann’s MJML plugin for Jetbrains editors to help me write, which makes things even easier, since it let’s me preview my work live.

I took one of their pretty templates, Recast and then stripped it down to meet my needs and match it as closely as I could to my old customised Antwort template.

Now all I do, is write, copy the HTML into the sendy editor and send. Life is good!

Feedback on this post?
Mail me at feedback at this domain.

P.S. Subscribe to my mailing list!

Or does revealjs work in the same vein as mjml? ↩︎

Kubernetes & CNCF Awards

feedback@janusworx.com (Mario Jason Braganza) — Mon, 20 Apr 2026 11:15:00 +0530

Intended Audience

Mostly me.

Click the image to embiggen!

This is a note of gratitude.
A tad late. But better late than never :)

I decided to get involved with the Kubernetes community on a lark about four years ago.
I wanted to learn about containers and orchestration, and I thought learning via joining the upstream community was it.

I was wrong 😂
Life got in the way and I only finally got around to learning it all last December.

But the community was interesting and helpful.
I got to learn the terms of art, by osmosis.
I got to see how a large community functions, how a large software project was run.

And in time, I got to pitch in and help a bit as well.
I pitched in with the Release Team for v1.25 to gain a sense of how a release comes about.
And a couple of years ago, I saw my good friend (and frequent collaborator) Priyanka buckling a bit under the weight of the work, in the various roles she had assumed to help out in the community. I asked her if there was anyway I could help her, that did not involve meetings or talking to people 😂¹ And she asked me if I could pitch in and help with helping folks gain membership into various Kubernetes orgs on Github.

I ayed.

She taught me the ropes. And I became a Github New Membership Coordinator for the Kubernetes project.

And all of this history to say, I’m grateful to Priyanka for this serendipitous opportunity.
I fell in love with the work.
For someone who has had to face a lot of cynicism in his real world day job, it’s been a lot of joy, welcoming folks into the project. It’s one of the happier parts of my day, when I sit at the desk to process the requests that come in.

I’m also grateful to the Kubernetes as well as the overarching CNCF communities for recognising what I do, for two successive years.
I was the recipient of the Kubernetes Contributor award in 2024 and the CNCF’s, Chop Wood Carry Water award in 2025.
I’m both, honoured and humbled.
I hope to contribute to this community, a while longer :)

Feedback on this post?
Mail me at feedback at this domain.

P.S. Subscribe to my mailing list!

I was well and truly burned out then ↩︎

Trying a Different Tack to Writing Work Posts

feedback@janusworx.com (Mario Jason Braganza) — Mon, 20 Apr 2026 11:00:00 +0530

Intended Audience

My tiny audience, who reads these posts via email or rss

My last post in this section / category was in December.
There ought to be more, because I learn so much every week.
Even in December, the massive move I did, needs documenting. (at least for me, if not for your reading)

So the new plan is to cut the doing and the learning in half (or by a third). And use that time to write about what I learned.
I know this about myself. I write real good … sometimes. But to even get to that point, I have to write a lot of drivel.

So, in essence, welcome to the drivel :)
The posts will get more frequent, so if that bothers you or I write about things that aren’t what you came to read me for, I won’t hold it against you if you choose to unsubscribe. I need to be able to write for me. And this is the only place where I can.

If you do decide to stay and read, as always, I’m grateful. I truly am.

Feedback on this post?
Mail me at feedback at this domain.

P.S. Subscribe to my mailing list!

The Big Plan: Change My VM to Be Gitops Driven

feedback@janusworx.com (Mario Jason Braganza) — Fri, 19 Dec 2025 18:00:00 +0530

Intended Audience

Me!

Update 2025-12-19: All done!

I just finished a move from one Hetzner VM to another.
The type of VMs are the same, in fact.
It’s just that the new VM and all the software on it are entirely software driven. I kept logging my progress in my notes, copy pasting the plan from day to day and ticking things off.
Now that it’s done and I still want to refer to it regularly, as the rest of the services come over, I wanted a place to keep it. And so this post, it is.

List of services that absolutely need to come over. Miscellaneous stuff later.
#

the main domain
french version of the website
the mastodon archive
the email distribution list
miniflux for rss feeds
joplin
baikal
~~discourse~~ (no more discourse!)
markdown editor (hedgedoc)
anki
huginn
syncthing
IRC: theLounge + znc (see if we can make do with a single service now (2025-12-15: we could!))
kanboard
Certs, Move them over, or figure out a way to generate and renew them via Ansible

The Big Point of The Big Plan
#

Save time and energy. Managing all the disparate services I use is taking more and more of my time. I need to claim that back, while being able to use said services.
Be gitops driven. Managing stuff gets easier. Tearing down things and rebuilding them gets easier
Have most everything I use, be in a Kubernetes cluster.
Be pragmatic enough to know that everything cannot be in a Kubernetes cluster and will have to live in the root VM
Have Flux CD manage everything in the cluster
Have Ansible Pull manage everything in the VM, acting as my single node. The point of doing this is not idempotency, rather to have everything in code; something that I can comment and uncomment and manipulate at will, something I can update at will and something that is documented. Never again will Future Jason have to scratch his head about, just how to go about doing something. (Long term note to self: Have the discipline to write tasks and drive everything with Ansible, despite the ease of “just doing it at the command line”)

The Big Plan (Done! 🎉🎉🎉)
#

The plan is to redo the cluster again and do my own instance of
- K3s
- Sealed Secrets
- Flux CD
- ~~Certmanager~~ (Not using it)
- ~~Letsencrypt~~ (using pre existing Letsencrypt certs)
- Get Traefik Ingress to work
- Figure out a way to get certs automatically into the cluster
And once that is done, figure out an app to move (Miniflux or Hedgedoc?); 2025-12-03: Kanboard it is!
Begin by moving (lifting and shifting in popular parlance) Kanboard to the cluster
- Cert will probably be needed (Wildcard cert works now, just like it does without the cluster)
- Convert a docker-compose to kubernetes manifests
- Learn how to configure an app with code
- Learn how to store data and back it up
- Figure out secrets, if there are any (for now sealed secrets ok, figure out vault and vault injection later)
- Learn how to tunnel through and reverse proxy
- Make Kubernetes manifests work with flux
- Figure out how to automate deployment of manual manifests
- Figure out how to migrate there if there is any in an old app
- Figue out how to automate updation of images in manual manifests
- Get another app (Miniflux) deployed
- Figure out what needs to happen as part of the lifecycle. What you want in the cluster, what stays out, do they intersect, how do updates of cluster happen? VM (node) updates as well?
- Then begin to think along the lines of Live Deploys. Prototype locally and once it works, migrate to production immediately
- Convert Kubernetes manifests to Helm Charts (optional, based on energy)
Go live! Git is source of truth. Two repos.
- One for the Main node and its update
  - Terraform will provision node and install package, setup firewall
  - Figure out how to get Terraform to get the node talking to the git forge
  - Structure repo, copy every thing node related there, and make sure stuff gets updated periodically and if possible, idempotantly, via ansible pull and a systemd timer
- The other one for k3s and flux
  - Convert everything I have done locally to run on prod. Add more steps as you do them below

Unrelated. Long term. Optional. Just here so that I remember
#

Get Moi publish script running
Redo Huginn Scenarios

Feedback on this post?
Mail me at feedback at this domain.

P.S. Subscribe to my mailing list!
Forward these posts and letters to your friends and get them to subscribe!

Moved From Reeder to Readkit

feedback@janusworx.com (Mario Jason Braganza) — Mon, 17 Nov 2025 08:59:16 +0530

I used Reeder for reading all my RSS feeds ever since the app launched over fifteen years ago. RSS is how I keep up with everything in the world outside and having a pleasant reading experience is very important to me.

It no longer serves that need for me.
For a long time, the way the app worked aligned with the way I want to read. Earlier this year though, the developer decided to take the app in a different direction, with a different reading experience. The older app is still available as Reeder Classic, but only a few months of use have shown me that the app is basically abandoned. The attention to detail is obviously now being applied to the new app.

Enter ReadKit.
I had used it briefly during the Google Reader apocalypse when every feed reader was scrambling to find new backends to sync to. Reeder similarly had switched to local only mode and was taking a while before it supported other services.
ReadKit in the meanwhile already had support for Feedwrangler, and so I switched to it until Reeder came back to speed.

And I’ve switched to it for the foreseeable future.
It looks beautiful!
It does everything I want, shows everything the way and want and behaves just the way I want it to. The only knock I have against it, is that it does not feel as fluid as Reeder does. But that’s nothing compared to the constant launch and relauch dance, I have to do with Reeder nowadays. Consistency and stability matter a lot to me.
Even better, it syncs natively with Miniflux, the service I use to actually fetch and read RSS feeds on my Linux desktop. No more Google Reader API!

This is a list of all my categories (with one of them expanded, click for a larger view)

and this is a list of unread articles in a feed, alongside one that is open (once again, click to enlarge if you want to see details)

That gnawing feeling has now gone away from the back of my brain.
The experience of reading and catching up with the world is once again a glorious experience thanks to ReadKit.

Feedback on this post?
Mail me at feedback at this domain or [continue the discourse here][2].

P.S. Subscribe to my mailing list!
Forward these posts and letters to your friends and get them to subscribe!

Kubernetes, Note to Self: Need Load Balancer Installed on Bare Metal

feedback@janusworx.com (Mario Jason Braganza) — Tue, 11 Nov 2025 07:37:54 +0530

Intended Audience

Mostly me. Also other grizzly sysadmins who are learning devops like me.

One thing that bit me when I was trying to expose my apps to the world when working on the home cluster, is that Kubernetes on bare metal—I was using Kind at the time—expects to talk to a load balancer service, which then talk to an actual load balancer. Which if you are using bare metal, you won’t usually have.
I had to then go expose a “NodePort” to gain access from outside.

So to expose my stuff in as “real world” a way as possible, we need to:

Either install a load balancer implementation like MetalLB. OR
Use a Kubernetes distribution that has a load balancer implementation built-in, like K3s.

I chose option 2 and used K3s, because I am, as they say in popular parlance, using Kubernetes at the edge.¹
In which case, I prefer to have as many batteries built-in as possible.

Although, from the articles I’ve read, if you’re doing a multiple node cluster, then you’re better off using MetalLB. ↩︎

K3s: Move Data Folder

feedback@janusworx.com (Mario Jason Braganza) — Mon, 10 Nov 2025 16:50:45 +0530

Intended Audience

Mostly me. Also other grizzly sysadmins who are learning devops like me.

Problem
#

My main SSD is too small to hold all my kubernetes and K3s shenanigans. (It’s only about 512gb) So I need a way to move my K3s data folder out to my big HDD.

Solution
#

Caution

Before you do any of the following, make sure to stop the cluster with:
sudo systemctl stop k3s

K3s supports having a custom data directory, by letting me specify a data-dir.
I decided to specify it in the K3s config file at /etc/rancher/k3s/config.
If it isn’t there (as it wasn’t in my case), just create one. You’ll need to do this as the root user.
This is what I now have in there.

data-dir: /BIGHDD/rancher/k3s

And then, I …

Created a rancher directory in my big hard disk. (root owns it)
Ran rsync as the root user: rsync -a /var/lib/rancher/k3s/ /BIGHDD/rancher/k3s/
Started up my cluster again with a sudo systemctl start k3s

Hugo Redirect From Post in One Section to Another

feedback@janusworx.com (Mario Jason Braganza) — Thu, 06 Nov 2025 08:34:36 +0530

Problem
#

I have a post in one section of my site, say notes and miscellanea
That is my canonical post.
But I also need it to show up in the work section of my site, possibly with a another name.
A Hugo alias in the yaml front matter does not work, since both sections exist and Hugo gets confused.

Solution
#

Make sure layouts/redirect/single.html exists within your Hugo folder structure and it contains just this line …

{{- template "alias.html" (dict "Permalink" .Params.target) -}}

Create your original post as usual
Then go create a post in your other¹ section with the following metadata:

---
title: <Your title here>
#date autogenerated by my archetype. if missing, add date
date: 2025-11-06T08:30:54+05:30 
type: redirect
target: <url-to-redirect-to>
summary: <Optional text, that will appear in the sectiones listing page>
---

The date field is critical. The post might not appear where you expect it, if it’s absent.
If not date, then you might need to use weight. One of the two is needed.
In my case, the metadata looked like this:

---
title: "The Plan for My New Hetzner VM"
date: 2025-11-06T08:30:54+05:30
type: redirect
target: /nm/2025-11-06/
summary: Alone and helpless, like you've lost your fight 
---

Done! You can see an entry for “The Plan for My New Hetzner VM” in my work section as you see in the pic below. Clicking that will redirect to the actual post in my Notes and Miscellanea section!

Tada 🎉

in my case, work ↩︎

The Plan for My New Hetzner VM

feedback@janusworx.com (Mario Jason Braganza) — Thu, 06 Nov 2025 08:30:54 +0530

Update: 2025-12-11
#

I got it done!

Preamble
#

If this post is weird and stream of consciousness like, that’s by design.
You are seeing a notes and miscellanea post, in the work category.
I want to be able to think through what I am doing. Mostly to clarify, document and explicitly state what I want. To avoid shifting goalposts (or rather to shift them intentionally, if need be.)
I’ll be updating and revising this post as I think and learn and experience things

Goals
#

Think about and write what I want to get done with my hetzner
For now, the how does not matter

What I want
#

A VM running K3s, that I can deploy my Hugo website to, as well all the apps that I choose to run on Janusworx
My main need with a cluster is the GitOps aspect. Not the scaling or redundancy. I want infrastructure as code and automated updates, as much as possible.
Everything goes through my Forgejo instance at home
I write a post and it should deploy to the website somehow
Apps that I put in there, either just my config files of already published images/deployments, or my own custom things, ought to be pushed to or pulled by the cluster and deployed. Upgrades should be a matter of manipulating code
Secrets should be self hosted, via environment variables or some self hosted solution. Never at a third party service

2025-11-06 09:35
#

Got tangled with Hugo in the morning.
I think I got the basics down, the VM comes up. K3s in installed, but in what order do I move my apps over?
Think I’ll the root domail last. And try beginning with the “simpler” apps first
Begin with familiarising myself with Helm
2025-11-06 10:06: Tea break
2025-11-06 13:06: Lunch break

2025-11-06 16:00
#

Back. Trying to port my dev terraform setup to production. It should just work. But it does not.
2025-11-06 16:59: Solved. I had somehow borked my Terraform state. Nuked it. And then things started working again
Show stopper now. My GPG key has expired. Need to figure out how to sign my git commits with SSH now
2025-11-06 18:26: I think I’ve licked it. Both Github and Forgejo don’t seem to be complaining. Let’s see what else I’ve broken over the coming days
2025-11-06 20:00: Done! My production cluster is up! And I can tear it down and build it up consistently. Now to work with it tomorrow.

Feedback on this post?
Mail me at feedback at this domain.

Make a Ruckus: 001, Done! Or Are We? Thank you, Dave!

feedback@janusworx.com (Mario Jason Braganza) — Sat, 20 Sep 2025 13:39:01 +0530

Cute wolf Mabel, who caused quite a Ruckus, courtesy David Beazley

I had started this series, to catalogue my journey learning Rust whilst doing David Beazley’s, Ruckus course.
And just like that, it’s over! 😂 The course is done, and I am already missing Dave and my fellow journey folk, Rodica, Yadu, Marko, Upul, Gabriela, and Eugene!
I never thought the end would be so bittersweet.

Worth it financially?
#

Yes! Oh, dear God, yes!
To begin with, I never thought that Dave’s courses were expensive. Yes, they cost money, but he’s given so much away for free, that I was willing to save up and put my money where my mouth was.
I also think, I’ve figured out Dave’s schtick with the idea of what he teaches.
Take something, something hard, (Write a compiler? Figure out distributed consensus in networked systems? The big ideas and paradigms in computing?); something that would take an average Joe, 4-6 months (or a college semester) to learn, and then compress it down to anything from a couple of days to a work week. Compress the time that is … not the knowledge.
Having that much knowledge dowloaded into your head in that short a time frame, is what (to use a Dave-ism) makes your brain explode.
And now having actually done it, I’m now swinging around to the price is just right or “He actually undercharges a bit” camp.
This is bloody addicting. I’m just going to save up and do them all as and when I can over time. Like Neo answering Tank’s, “How about some more?”, with a Hell yes!

Worth it for the knowledge I gained?
#

Have I already said, Oh, dear God, Yes?! Doubly so, here!
Learning it myself, I realised that Rust had paradigms, teasing me at the edge of my vision, which I couldn’t quite see, frustrating my attempts to learn the language.
Python (C/OOP) is my ur-language. Having that, made it easy to learn Go. Rust turned out to be a different beast altogether. Having the big ideas of Rust laid bare and then explained thoroughly was therefore, such an eye opening experience. And I learned it while doing something else, that is inherently hard. Dave makes it easy though, like I describe below.

I just wanted to learn Rust. I didn’t really want to write a small programming language.
But doing that, made me learn Rust without even realising it. He’s that good. The pedagogical method is that good.
Most big ideas were in there. Ownership and borrowing, check. Compare and contrast how things are done in other languages, double check. Macro magic, check and triple check.
The course was more focussed on the why of Rust, more than the more easily discoverable, how.
A fascinating mental model was looking at Rust like Earth. An inner strict core, surrounded by a more relaxed, friendlier outer shell.

Daily work was tremendously, mentally taxing. In a good way.
Dave keeps saying that minds explode due to the material. But no good Sir, that is not true!
To Dave I say, “’Tis thou, Dave, who dost cause the explosions of brains!”
He starts with an innocent problem, something seemingly innocuous, and then lays his trap.
Simple solutions are shown not to work, teensy yet ghastly exceptions show up to trip you, dreadful scenarios that you’d never imagine, lie beneath the surface. And your mind gets wound up so tight, that when Dave finally has had his fun, takes pity on us and shows us the solution, it naturally goes, Boom!

I cannot wait, to have my brains exploded once again :)

Am I done with this series?
#

No.
I jumped into the deep end of the pool, by attempting to learn Rust with a hard project at the same time.
So while I’ve learned a lot, there’s plenty of work to do.
While the others could do their programming exercises quickly, I have seriously lagged behind.
With the notes I’ve made and the resources that I have access to, I will attempt to redo this week myself over the coming months and write about it, and ponder, struggle, work through it here.

Dave as teacher and all round good egg
#

The man puts in tremendous amounts of work.
Every bit of the course, is planned and structured meticulously and then he just riffs off of the ideas.
There was a collaborative Github repo ready and waiting, there were video recordings ready as soon as the day was done.
Dave never pretends to have all the answers. 99% of the time, he does and hides it well, so we do the thinking, but the few times he didn’t, he let us know and then came back to us with answers and thoughts either in the next session or the next day.
Unfailingly polite, kind and generous to a fault, he fostered an extremely collaborative environment, in which, even when there were other folks much more learned and experienced than I; I never felt like a dunce. I never felt spoken down to.
All my kooky ideas were evaluated, and then Dave would point out why doing something like what I suggested wouldn’t quite work or why stuff would crash and burn or why what I suggested doing eeny, meeny, miny, moe was in fact, pretty good.
I learned not just what, not just how, but also why.
And I learned it all in a place that was nice, and friendly, and warm, and kind.

Dave, thank you for all you do! I’m very grateful!

Make a Ruckus: 000, The Motivation

feedback@janusworx.com (Mario Jason Braganza) — Thu, 11 Sep 2025 13:47:52 +0530

Wolf howl, by Lorc, CC BY 3.0, via Wikimedia Commons

I signed up for the latest cohort of David Beazley’s, Ruckus.
The basic idea, behind Ruckus is to write a small interpreter in Rust!

Why?
#

From the Ruckus page:

This course tackles Rust by stepping back and thinking about it from the standpoint of a “programming languages” project. If Rust is really so different, how would you go about finding out how? Suffice to say, building a programming language will cut right to the heart of the matter.

This is exactly how I learned Python. Only it took me an inordinately long time. And I didn’t even realise that I was doing it that way. While most of my friends were happy just using Python, I could not quite do it right, because I had no fundamentals¹, and since my background has been tech support, hardware and people management, I needed to unlearn a ton of stuff.

I’m cautiously hoping Ruckus will help me do the same with Rust, only in a turbo charged manner.

Rust fascinates me. And I quite like my journey to its doorstep.
I learned writing for myself with Python. But when it came time to sharing my tiny projects with friends and family, I was a bit stumped.
I learned Go, just so I could give them a binary and then instruct them with a “just run this.”
And while that solved my distribution problem, Go’s type system made me realise, just how well suited a typed language is for collaboration.
Which brings me to Rust. Long term, this is the language², I would love to write for performant code, because Rust’s big picture features (safe, performant, easy to distribute) are exactly what I want for all my tiny bits and bobs.

Which in turn brings me back to Ruckus.
Getting a teensy peek at how everything works underneath, while at the same time, getting a firm handle on the language seems like a fun challenge.
I also wanted to try what Cal Newport, calls a grand gesture in Deep Work:

The concept is simple: By leveraging a radical change to your normal environment, coupled perhaps with a significant investment of effort or money, all dedicated toward supporting a deep work task, you increase the perceived importance of the task. This boost in importance reduces your mind’s instinct to procrastinate and delivers an injection of motivation and energy.
[…] The dominant force is the psychology of committing so seriously to the task at hand. To put yourself in an exotic location to focus on a writing project, or to take a week off from work just to think, or to lock yourself in a hotel room until you complete an important invention: These gestures push your deep goal to a level of mental priority that helps unlock the needed mental resources. Sometimes to go deep, you must first go big.

Make no mistake, my current state is that of an enthusiastic, ignorant newbie. How hard can it be? I’m woefully unprepared for what might come. The only thing I know, is that I don’t know. I fully expect to be like the wolf above, losing my mind, howling at the moon and generally creating a ruckus.

Feedback on this post?
Mail me at feedback at this domain or continue the discourse here.

P.S. Subscribe to my mailing list!
Forward these posts and letters to your friends and get them to subscribe!

I have some now ↩︎
I thought I could be pragmatic about the giant that stewards Go, but I realise that it’s something that’ll always stick in my craw. I have no issues using it with other folks or in a collaborative project. I just don’t want to use it for myself and my projects. ↩︎

Emacs Package Updation Checklist

feedback@janusworx.com (Mario Jason Braganza) — Sun, 06 Jul 2025 08:36:44 +0530

I’ve never updated my Emacs packages until recently, because Emacs is where all my writing happens, and so I’m justifiably paranoid.
But then some packages stopped working, due to various circumstances¹ and an update solved it.

So I’ve decided to update my packages once a quarter, so that I don’t lose days yak shaving when something goes wrong and I handle breakage on my terms and not the machine’s.

As far as package management goes, I want to keep things simple.
In fact, I still haven’t graduated to use-package or straight.el because my package needs are few and conservative². And so, while there are automatic update options out there, I’ll just stick to updating them manually, every quarter.

Ergo, this is the checklist I’ll use next time onwards …

Stop emacs user service, systemctl --user stop emacs
Backup emacs folder in ~/.config
Start emacs manually (not the service).
M-x package-refresh-contents
M-x package-upgrade-all
Problems? Quit emacs. Revert backup folder.
In the end, start emacs user sevice, systemctl --user start emacs

There’s an Org mode task, scheduled quarterly, so that I won’t forget.

Feedback on this post?
Mail me at feedback at this domain or [continue the discourse here][3].

P.S. Subscribe to my mailing list!
Forward these posts and letters to your friends and get them to subscribe!
P.P.S. Feed my insatiable reading habit.

While I don’t want updated packages, I do want updated Emacs and that broke stuff 😂 ↩︎
The biggest change I forsee, is if Jetbrains ever turn evil and I have to move off their editors and subsequently need to use Emacs as an IDE ↩︎

Forgejo Workflow Action to Publish Hugo Blog

feedback@janusworx.com (Mario Jason Braganza) — Mon, 09 Jun 2025 15:09:16 +0530

I’d gotten this working earlier and wanted to post it up.
And then promptly forgot about it.
Until today that is, when I wanted to make changes to it.

Here it is for posterity.
The newest change is that the action now runs every six hours.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54


name: Build and Deploy Janusworx website
on:
 schedule:
 - cron: '0 */6 * * *'
 workflow_dispatch:
 push:
 branches:
 - main

jobs:
 build-and-deploy:
 runs-on: intel # make sure there are runners tagged "intel" on your network or replace this with something of your own

 env:
 HUGO_VERSION: ${{ vars.HUGO_VERSION }}
 steps: 
 - name: Install wget & rsync
 run: apt-get update && apt-get install wget rsync -y

 - name: Install Hugo
 run: |
 wget -O ${{ runner.temp }}/hugo.deb https://github.com/gohugoio/hugo/releases/download/v${HUGO_VERSION}/hugo_extended_${HUGO_VERSION}_linux-amd64.deb \
 && dpkg -i ${{ runner.temp }}/hugo.deb
 
 - name: Checkout the website repo
 uses: actions/checkout@v4
 with:
 submodules: recursive
 fetch-depth: 0
 
 - name: Build website with Hugo
 env:
 HUGO_ENVIRONMENT: production
 TZ: Asia/Kolkata
 run: |
 hugo \
 --gc \
 --minify
 
 - name: Install SSH key
 run: |
 mkdir -p ~/.ssh
 echo "${{ secrets.SSH_PRIVATE_KEY }}" > ~/.ssh/id_ed25519
 chmod 600 ~/.ssh/id_ed25519
 echo "${{ secrets.SSH_KNOWN_HOSTS }}" > ~/.ssh/known_hosts
 chmod 644 ~/.ssh/known_hosts

 - name: Deploy website with rsync
 env:
 REMOTE_HOST: ${{ vars.REMOTE_HOST }}
 REMOTE_USER: ${{ vars.REMOTE_USER }}
 REMOTE_PATH: ${{ vars.REMOTE_PATH }}
 run: |
 rsync -avz --delete public/ $REMOTE_USER@$REMOTE_HOST:$REMOTE_PATH

Notes:
It needs a few variables set, in the repo (repo->Settings->Actions->Variables), which are the details I need (Hugo version, Remote Host IP, Remote path to copy the built output to, and the Remote user to use for SSH) to push via rsync to the host.
Hugo version is in there, because Hugo breaks things all the time, so I upgrade cautiously.

It also needs a couple of secrets, set in the repo (repo->Settings->Actions->Secrets) to enable rsync to copy. (SSH_KNOWN_HOSTS and SSH_PRIVATE_KEY). Also best if this is a dedicated private key just for copying. Running ssh-keyscan "your-host-name-or-ip" > known_host_data.txt should dump the data you’d need for SSH_KNOWN_HOSTS

With the changes I made today, the workflow can publish

on demand
every 6 hours
every time I commit to the repo

I added time based publishing, so that I could publish future dated posts as and when the time rolls around.
With this, I think I’m done tinkering with my blog, with the exception of moving it out of my home network and out on to my VM, so that I can truly publish from anywhere. That is a big heft for me currently and a job for future Jason.

Here’s What is Stopping Us From Using Free Software

feedback@janusworx.com (Mario Jason Braganza) — Tue, 18 Feb 2025 12:13:45 +0530

A couple of thoughts on Robin’s post.
Because I’ve spent a lifetime using both FOSS and proprietary software and more than two decades, advising and supporting people and businesses who do the same.

To paraphrase Robin’s own point. Most people don’t care.¹

If it runs Minetest, then it’s fine.

aka “If it does the thing, I want it to do, without too much cognitive overhead, I’ll use it.”
Learning the skill, so I can do this task, is definitely not what the world at large wants to do. Nor do I imagine, do most folk have the time and the bandwidth. iTunes (and Netflix) when they launched², were the prime examples of this. Convenience and ease of use trumps Free. Also by that same token, VLC is probably one of the most installed and used video players in the world, not because they are FOSS, but because they make it easy to install and use.

Some people feel strongly about having control over what they do, over their data. Those folks were those I found easiest to move to Free software, because it aligned with what they wanted.
Support costs are about the same.
If a small business chooses to use free software, the cost of supporting it, if stuff breaks, (specially if it’s business critical³) probably is about the same as paying for the proprietary version. The fact that paying and getting it fixed benefits the world at large is lost on most. Or maybe it’s something, we Free Software evangelists could learn to to better. Robin promises his team, first-level support, but what happens when he goes on to better opportunities and is no longer around?

So I agree with most of what Robin says, with just one caveat.

People aren’t afraid of change. Atleast not the kind of change, Free Software poses.
People just don’t have the time and the energy to deal with change, with everything else happening in their lives.
Free software is just not as important to them, as it is to us. That is all.⁴

And I’d venture to say, that’s perfectly ok. ↩︎
before they went to pot ↩︎
rpm updates, bork my expensive garment printing machine, anyone? ↩︎
And once again, that’s perfectly ok too. ↩︎

Blog Questions Challenge 2025

feedback@janusworx.com (Mario Jason Braganza) — Tue, 04 Feb 2025 12:10:17 +0530

Ava started this, Kev modified this, and Saptak egged me on to write this. So here goes …

What the home page looks like in 2025

1. Why did you make the blog in the first place?
#

I write for me mostly. Because writing helps me think. My thoughts are too scattered otherwise. I can’t not write. I’ve always written. Privately, publicly, there’s always been some place where I’ve jotted things down.

2. What platform are you using to manage your blog and why did you choose it?
#

I use Hugo to generate the site, which I host on my own Hetzner VM. I use it because I outgrew my previous tool, Nikola, which still holds a dear place in my heart. While Hugo is enormously complex, it is also deceptively simple enough to get started with. And it’s fast. That’s what I love about it. It lets me write. It does not get in my way. It lets me preview what I’m doing with its live server. And it’s unbelievably fast.

3. Have you blogged on other platforms before?
#

I’ve been writing in some form or other since the late 90s. So … yea :)
Livejournal, Blogger, self hosted Wordpress, wordpress.com, Posterous, Tumblr, self hosted Wordpress, self hosted Ghost, Nikola and now Hugo. It’s been quite a ride!

4. How do you write your posts?
#

I write them in Emacs (in Markdown, using Markdown Mode) on my desktop, with Hugo Server running alongside giving me a preview of what things will look like. Once I commit it to my self hosted Forgejo instance, an action publishes the site automatically

5. When do you feel most inspired to write?
#

I never do. I write because it helps me function. And yet, it always feels like a chore.

6. Do you publish immediately after writing or do you let it simmer a bit as a draft?
#

I always publish it immediately. I almost never write something that is deeply thought out, that needs to stand the test of time. It’s the process, the writing, the putting words out of my mind, through my fingers down to paper, that leads to the result, the thought, the opinion, the aha, the insight. So I’m never quite done. Which means if I ever wait for finished, the post will never get published. The moment I publish something, is invariably the moment something needs changing. So I just go back and edit it. I never notify folks about updating things on the first day. If I ever edit something much later than a day or two, then I do.

7. Your favorite post on your blog?
#

None. All. They’re my thoughts, so depending on my mood, they’re either worthless or priceless gems!

8. Any future plans for your blog? Maybe a redesign, changing the tag system, etc.?
#

Not really. For all my wandering, I’ve only ever moved when my tools outgrew me¹, or I outgrew my tools. For now, Hugo does all I ask of it, without getting in the way. The day that changes, will be the day I move.

I’ll ask Priyanka, Sreeram, Sandeep, Pradhvan, Rahul, Bhavin, Elia, Mandar, Saptak, Farhaan, Robin and Kushal to share more, if they have the time, energy and the inclination.

Folks, who’ve answered my call to arms! Go read their answers too!

Robin
Elia

Wordpress.com too restrictive, Ghost stopped serving my specific needs etc. ↩︎

Hack Hugo Post Metadata With Python

feedback@janusworx.com (Mario Jason Braganza) — Mon, 03 Feb 2025 09:33:37 +0530

A while back, I rejigged the sections on my site to better reflect how I think and write.
Which meant, all the urls, on all my posts changed, since they now used the new category as a slug, instead of ye old /blog.

For e.g. https://janusworx.com/blog/using-hugo-variables-to-help-with-mailto-links-in-hugo/ …
was now at, https://janusworx.com/work/using-hugo-variables-to-help-with-mailto-links-in-hugo/

After searching a bit, I found that Hugo supported aliases. For me, it would redirect the original /blog path url to its new location
Ass I had to do, was add an aliases: ["/blog/old-post-slug"] line to each post’s metadata.¹
Line 4 in the snippet below shows, what I added to fix the post above.

1
2
3
4
5
6
7
8


---
title: "Using Hugo Variables to Help With Mailto Links in Hugo"
date: 2024-05-30T18:17:35+05:30
aliases: ["/blog/using-hugo-variables-to-help-with-mailto-links-in-hugo"]
categories: ["work"]
tags: [100WordHabit, Dgplug, Hugo]
summary: Shortcodes! Hugo Variables in Shortcodes!
---

I did not want to do this by hand for 800+ posts.
One stroke of luck for me, was that I let Hugo use its default behaviour of generating url slugs from the file names. So even if the category slugs had changed (from /blog to /work or from /blog to /personal), the url slugs would stay the same. Which meant, I could whip up a script to run through all my markdown posts and add the alias line.
So I did.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12


from pathlib import Path

INPUT_FOLDER = Path("old-posts-folder")
OUTPUT_FOLDER = Path("modified-posts-folder")

for each_file in INPUT_FOLDER.iterdir():
 with open(each_file, 'r') as file_to_read:
 alias_derived_from_file = each_file.stem
 contents_as_list = file_to_read.readlines()
 contents_as_list.insert(3, f"aliases: [\"/blog/{alias_derived_from_file}\"]\n")
 with open(Path(OUTPUT_FOLDER, each_file.name), 'w+') as file_to_write:
 file_to_write.writelines(contents_as_list)

It takes all the posts from my old folder, inserts the alias line and puts them into a new folder.² In essence, take each file, figure out the url slug from the file name, read in the contents as a list, insert my alias at position 3 (fourth actually. zero based indexing) of the list (below the title and date) and then write it all out to a new file.

I ran it, published the site and then went to check on the old urls with bated breath.
Hurrah, it all worked :)

it’s a list, so I can add more aliases if I want to ↩︎
no sense in botching up my originals :) ↩︎

Tmux Start Session Maximized With Three Panes

feedback@janusworx.com (Mario Jason Braganza) — Sat, 01 Feb 2025 09:27:25 +0530

So I got tired of starting up a dedicated tmux session to manage all the work related to my writing sessions. Over the past few months, I’ve boiled it down to three.
And it still irks me that I have to …

Launch Terminal
Launch Tmux
Split it into three windows err … panes.
Go to the top left pane and launch Hugo server
Switch to the right pane and then launch Emacs with whatever new post I want to write today.

So of course, a Rube Goldberg-esque, tiny bash-pipey monster took form. It now resides, chained to an alias, hssx¹ in my .bash_aliases file

1
2
3


alias hssx='cd /path/to/my/hugo/folder && \
xdotool windowsize $(xdotool getactivewindow) 100% 100% && \
tmux new-session \; split-window -h \; select-pane -l \; split-window -v \; select-pane -U \; send-keys "hugo serve" C-m \; select-pane -R'

The first line switches to my hugo folder
The second calls xdotool and maximises the terminal window
And the last line is a series of instructions to the tmux command. I’ve split it below for readability.

tmux new-session \; #Tmux start a new session
split-window -h \; # Split the window into two vertical panes
select-pane -l \; # Switch to the left pane
split-window -v \; # Split that into two horizontal panes
select-pane -U \; # Select the upper pane
send-keys "hugo serve" C-m \; # Type in `hugo serve` followed by Enter
select-pane -R # Select the right pane

And boom!

Click pic for a larger version

hugo start session and x just because all my aliases have ended with x for years and years ↩︎

View Only the Day, With Org Agenda

feedback@janusworx.com (Mario Jason Braganza) — Mon, 20 Jan 2025 08:50:40 +0530

Org Mode, by default, shows me the whole week, when I pull up my agenda. That’s not how I work though. I normally look at the week on Monday mornings (or Sunday evenings) to plan out the week and then work everyday by just looking at what I ought to get done that particular day.

Like Dale Carnegie says,

Shut off the past! Let the dead past bury its dead. Shut out the yesterdays which have lighted fools the way to dusty death.

The load of tomorrow, added to that of yesterday, carried today, makes the strongest falter. Shut off the future as tightly as the past.

The future is today. There is no tomorrow. The day of man’s salvation is now.

Waste of energy, mental distress, nervous worries dog the steps of a man who is anxious about the future. Shut close, then the great fore and aft bulkheads, and prepare to cultivate the habit of life of day-tight compartments.

So I want then, my Org Mode to do show me, only the things, I have on my plate today.

And Org is nothing if not adaptable. org-agenda-span is the variable that controls what the Agends shows me. According to the docs,

This variable can be set to any number of days you want to see by default in the agenda, or to a span name, such a day, week, month or year.

So that’s what I did. This little snippet went into my init file

;; Set Agenda to show a day as the default timespan, instead of a week
(setq org-agenda-span 'day)

Et voilà!

More Espanso & Emacs Shenanigans

feedback@janusworx.com (Mario Jason Braganza) — Sun, 19 Jan 2025 12:45:09 +0530

Now that I use my local machine to draft all my posts, desktop Emacs makes everything easier … with one exception. It swallows up a lot of my Espanso expansions.

My current bugbear is that it will not render a = when expanding.
I know this is entirely something that I brought on myself, and neither Emacs or Espanso in isolation will have issues.

My workaround after rummaging through Espanso’s documentation was to set the backend to use xdotool, in my config. (~/.config/espanso/config/default.yml)

x11_use_xdotool_backend: true

One downside now is that it no longer jumps back to a predefined location in the snippets where I’ve defined such a thing. No matter. I use them rarely and I’ll figure something out later. (Alongside why desktop emacs won’t render markdown expansions sometimes as well)

For now, most of my expansions work. Hurrah!

Notes Now Has No RSS Feed

feedback@janusworx.com (Mario Jason Braganza) — Sat, 04 Jan 2025 16:34:20 +0530

Finally figured out how to disable RSS feeds on a post by post basis.

I needed to do this for my stream of consciousness / wip / progress notes in the Notes and Miscellanea section I did not want to disable the whole section, because some posts, I do want to leak out.

What it boiled down to was getting the RSS template to ignore sites with some parameter set in the front matter. Most posts on the web do a rss_disabled: true or something to that effect.

Joe Mooring over on the Hugo Discourse logically pointed out that I’d want such posts to be disabled in the sitemap as well, so I’d do well to just piggy back off of that.

So all I had to do was set the following in the my front matter

sitemap:
 disable: true

and then configure the embedded RSS template¹ to just range through pages that didn’t have that parameter set, like so …

{{- range where $pages "Sitemap.Disable" "ne" true }}

The original line would’ve been something like

{{- range $pages }}

You’ll need to create a new feed file by copying the contents of the embedded template to a file called rss.xml in the layouts/_default/ directory

Once that’s done, all should be hunky dory.
Posts with this frontmatter, will now no longer appear in RSS feeds or the sitemap.
To make things simpler, I edited this sections archetype, so that the next time I create a new Notes & Miscellenea post, all of this metadata is already set for me.

If you already have it defined in your theme, then override it. ↩︎

Fixing Espanso Expansions

feedback@janusworx.com (Mario Jason Braganza) — Sun, 29 Sep 2024 17:05:19 +0530

I never had the time to deal with my Espanso¹ hijinks until today.
While it worked perfectly, when I installed it all those years ago, when I migrated over from the Mac, Espanso itself has changed and evolved over the years.
It took over my old configuration like a champ and mostly worked, with the exception of a few shortcuts; ones that I frequently used 😂

Emacs was one application of mine that never quite worked right with Espanso.
I’d frequently get timed out waiting for reply from selection owner whenever I tried expansions in there. Typing :joy to get 😂 would work in every other program, but no joy with Emacs, in additon to plenty of other expansions err … not expanding.

All my browser url expansions would not expand properly either, with mangled expansions most of the time.

So today I dove in to the docs, and realised two things.
1. My emacs needed a longer time out
2. Espanso now tries to identify the kind of text, and maybe those were causing my issues?

Emacs
#

I realised I needed a longer clipboard threshold, only for Emacs.
So I created an App specific configuration, just for Emacs to use and gave it said option. Here’s what the contents of my espanso/config/emacs.yml look like

filter_class: 'Emacs'
clipboard_threshold: 10000

Rich Text Expansions
#

That helped with a lot of expansions in Emacs, but not with my joy expansion.
And not with stuff that were links and oh … links! and html! and markdown! Could those be the culprits?²
The docs mention that Espanso now has rich text support
What that means, is that the trigger nows supports two new keywords html and markdown in addition ye ole replace So I changed most of my affected shortcuts to either of those two keywords; markdown for most everything and html for linky stuff. Here’s what my beloved joy looks like now …

matches:
 - trigger: ":joy"
 markdown: 😂

And those two things did it! Every shortcut expands everwhere! What joy 😂

Espanso lets you assign short snippets of text to longer streams of text, that you’d frequently use. I use :adhx for example, and it expands out to type my whole address. Pretty handy! ↩︎
dun! dun! dun! they were! ↩︎

Testing Opengraph Logo

feedback@janusworx.com (Mario Jason Braganza) — Sat, 28 Sep 2024 17:46:14 +0530

‘Romeo, Romeo! Wherefore art thou Romeo?’

(Romeo and Juliet, Act 2, Scene 2)

Forgejo

feedback@janusworx.com (Mario Jason Braganza) — Sat, 28 Sep 2024 15:39:12 +0530

Setup a new instance of Forgejo for myself today.
The first thing I felt, was … instant relief.
I didn’t realise just how much cognitive discomfort I was feeling, because Github was the only place I had all my code. As well as mirrors/forks of all the stuff I loved.
I’m going to slowly move over to my little Forgejo instance as I learn more.

Right now I’ve set up pull mirrors for all the projects I love.

Next step is to figure out how to automate my Hugo blog deployment and more importantly; having all my posts in source control.

Smol Note to Self, on Deploying Stuff

feedback@janusworx.com (Mario Jason Braganza) — Sat, 28 Sep 2024 14:34:47 +0530

Anything I want folks to see and/or interact with, goes on the public VM
Anything I want to host for myself, goes on the Pi.

anki-push-u, Creating a Tiny Pushover Addon for Anki

feedback@janusworx.com (Mario Jason Braganza) — Sat, 21 Sep 2024 11:31:39 +0530

I want to slowly increase my French vocabulary, so I got this comprehensive frequency word deck from Shared Decks section of the Anki website.

I keep forgetting to look it up during the day, after my morning session and the only way I can get those stubborn words¹ to stick in my mind, is if I keep doing the deck 4–5 times a day.

I know! With all my newfound devops/python skills, could I figure out a way to remind myself to do it? Turns out I can! :)

I already use Pushover, to get notified of darn near anything.²
And I read enough of Anki’s add-on documentation to know I could whip something up.
So a bit of searching on the net, a bit of jiggery pokery with Claude, and some spelunking through Anki’s source code and forums later, I present to you … anki-push-u!³

As long as your Anki’s running, this little add-on will find cards due, at the interval you tell it to, and then notify you wherever you have Pushover running!
No more forgetting due cards!

Find the add-on and instructions over at Github.

Feedback on this post? Mail me at feedback at this domain

P.S. Subscribe to my mailing list!
Forward these posts and letters to your friends and get them to subscribe!
P.P.S. Feed my insatiable reading habit.

I’m looking at you, ailleurs ↩︎
Hugo updates on three machines, calibre updates on two machines, podcast downloads and quite a bit else ↩︎
It pushes you to do your due cards! Tada! ↩︎

Some Site Housekeeping (New Sections and RSS Feeds)

feedback@janusworx.com (Mario Jason Braganza) — Sat, 14 Sep 2024 11:36:16 +0530

Hard at work, on the site
Click the image, to embiggen

If you look at the site today, you’ll see that ye olde blog section has disappeared.
I’ve replaced it with three sections, Work, Personal and Reading.
I’ve done this primarily for me.

While I want all my writing in one place, I want them organised in piles that correspond with the three major modes of my brain.
Work to hold everything I consider work.
Reading for the bookworm in me.
And finally personal for everything else.

The problem (my problem, in my mind) with the old blog was too much mental whiplash when I read the blog. I would go from poetry to docker to antifragility to photos with every turn of the page.
I didn’t like it.
Ergo, the reorganisation.

Every one in their own pretty place now.
Click the image, to embiggen

So now, whenever I am browsing my posts in the work section, I see only tech, and computers and docker and programming and all that good stuff.
Same with the personal stuff and the reading.

How does this affect you, my kind and gentle reader?
#

If you are a subscriber to the email list, nothing changes. I know I’ve been lax with updates, but hopefully this reorg should break up the log jam and help me write more.
If you just come here to the site, once in a while to read through, hopefully, you’ll find it easy just like me, to read posts without too much cognitive whipsawing.
And finally, if you subscribe via RSS, you my friend are probably affected.
The main feed, with all articles, still remains the same, so if you drink from the firehose, nothing changes.
If you’re subscribed to a tag specific feed, like Emacs or DGPLUG, nothing changes.
The only change is that the blog feed has disappeared. In its place are three spanking new feeds, representing my new sections.
And that about does it. Happy Reading!

Notes to Self, Compiling Qbittorrent on a Raspberry Pi 5

feedback@janusworx.com (Mario Jason Braganza) — Mon, 19 Aug 2024 14:13:31 +0530

Update, 2024-08-20:
Added a missing step, to install the qBittorrent files to their relevant places
Also, it had to happen. No releases for close to three months and then just as I learn to compile, a new version has to release the very next day 😂
It gave me a chance to test my notes. It hardly took me a few minutes to get the new release in place. So all is good!

A compilation of letters, courtesy Cosmos

I wanted to always have the lastest stable version of qBittorent on my Pi.
So today, I sat down and followed the project’s excellent instructions and got it running.
I did however stray off the beaten path a teensy bit¹.
I didn’t want a gui and I wanted to be able to control it with systemd.

So here’s my checklist for next release. I’m currently running Bookworm on my Pi 5

In case you run into problems, compile instructions are here: https://github.com/qbittorrent/qBittorrent/wiki/Compilation-Debian,-Ubuntu,-and-derivatives
Do all off this after switching to root. (or sudo)
Install build tools
sudo apt install build-essential cmake git ninja-build pkg-config libboost-dev libssl-dev zlib1g-dev libgl1-mesa-dev
Install Qt dependencies (qtbase5-private-dev is not part of the instructions, but compile will fail if you don’t include it)²
sudo apt install --no-install-recommends qtbase5-dev qtbase5-private-dev qttools5-dev libqt5svg5-dev
Qt requires Python at runtime, so: sudo apt install python3
Install the libtorrent library. I compiled from source. Check page for fresh instructions, if stuff here doesn’t work
```
git clone --recurse-submodules https://github.com/arvidn/libtorrent.git
cd libtorrent
git checkout RC_2_0 # or a 2.0.x tag
cmake -B build -DCMAKE_BUILD_TYPE=RelWithDebInfo -DCMAKE_INSTALL_PREFIX=/usr/local
cmake --build build -j4 # the -j4 is to do on four cores at once. my addition.
sudo cmake --install build
```
If you ever want to uninstall libtorrent, then don’t delete this main folder. The build process would have created a install_manifest.txt in the build folder. So you can then uninstall with: sudo xargs rm < install_manifest.txt
Get qb source: https://www.qbittorrent.org/download.
Unzip: tar xzvf qbittorrent-4.6.5.tar.gz and then cd qbittorrent-4.6.5/
Create Build configuration (disabled GUI and enabled systemd. Those aren’t part of the defaults)
cmake -G "Ninja" -B build -DCMAKE_BUILD_TYPE=RelWithDebInfo -DCMAKE_INSTALL_PREFIX=/usr/local -DGUI=OFF -DSYSTEMD=ON
Build cmake --build build
Install all the files to their relevant locations cmake --install build

Install service to run for a specific user.

systemctl daemon-reload
systemctl enable qbittorrent-nox@<some-user>.service
systemctl start qbittorrent-nox@<some-user>.service

We’re done.

which the project does make provisions for. ↩︎
Target "qbt_base" links to: Qt::CorePrivate but the target was not found. or words to that effect ↩︎

Doing Devops, Dropping Rust

feedback@janusworx.com (Mario Jason Braganza) — Fri, 07 Jun 2024 16:37:50 +0530

Been struggling to do it all over the past few months.
But no! Life says no :)

So I’m building some margin into my life.
And the only thing I can talk about work wise, is the familiarity I am trying to gain as part of the career pivot.
The Devops Stuff.
The only thing that is career adjacent to all the stuff I did as a sysadmin and then consultant. So hopefully lots of boring stuff about exciting tech as I try to build new mental models, to fit over old ones.

Problem though, is what do I write?
I don’t have a good spin on most computer devops stuff. I don’t know what to say.
So for now, they probably will be, just like I point above, boring progress logs.
Mostly consistent, to give me accountability.
And hopefully, I’ll find a voice.

On the personal front, books, reading, art, language, poetry, photography and other posts will continue. Those are what make my soul sing :)

Note to Self, Linking to a Headline in an Org File

feedback@janusworx.com (Mario Jason Braganza) — Thu, 06 Jun 2024 17:31:37 +0530

I did one of those periodic, empty your mind sessions today.
Everything is now in a brain dump org file.
And now, like a every good productivity system preaches, I want to review the dump periodically.
There is a task for my miscellenea for e.g. Could I convert it to a link and have it open the file at that precise location?
Turns out Org Mode links let me do just that.
All I need to do, is first “store a link” in Org parlance. I do that by going to my target file that has the dump (the bottom file in the screenshot), go to the section that I need linking to (in my case, Misc Brain Dump), and hit C-c l

After which I go to the file which needs to contain my link, and select my text that I’ll want converted into a link (“inserting a link” in org speak. In my case, the top file in the screenshot below).
I select the Misc Brain Dump part of the “Look at the Misc Brain Dump” task and hit
C-c C-l
It then switches me to a buffer that shows the links that Org currently has stored, with the minibuffer prompting me to choose one

I use the arrow keys to choose the one I want, hit return and it creates the link for me.

Here, I’ve switched away from my brain dump. (I’ve closed the file in fact)

I click the link in my tasks file and …

It opens right up, at the exact location I want. Woohoo!

And just like everything else with Org, it’s plain, simple, elegant, text functionality.
I examined the link and it was a borg standard Org link, with the link having the header tacked on the end, like so …
file:brain-dump.org::*Misc Brain Dump¹

which should have been evident when I chose the link to insert, but I only saw it and went Aha! much later 😂 ↩︎

Moved My Music Library From Jellyfin to Emby Today

feedback@janusworx.com (Mario Jason Braganza) — Tue, 04 Jun 2024 10:47:23 +0530

I had moved off Plex, close to two years ago, because … reasons.
Unable to decide between Emby and Jellyfin and having just learned containers at the time, I deployed both on my Pi.
Emby handled video duties while Jellyfin did music.

And now after close to two years, I’m moving off Jellyfin to Emby.
The only reason being its iOS client.
It still has a lot of growing up to do, and I’ve reached the end of my tether trying to adjust to its foibles.
I cannot turn the phone off, else the music goes off.
I cannot switch away from the Jellyfin app, else the music goes off.
If I get a call, then the app sorta-kinda loses its mind and I have to kill it and relaunch it.

All I want from my music player is to be a good audio citizen on iOS. Connect to the server, read my songs, play my music. And do it in the background.

While I am sympathetic to the priorities of the open source Jellyfin team, I am not getting any younger and my hour long walks are the only time I get to enjoy my music.

So … I moved.

Setting the stage
#

There had to be two instances of Emby running. I don’t want my video online and so I don’t want it exposed
Video is secondary anyway. So that had to be the second container. The main one is Music.
Music needs to be accessible from anywhere on the web.

Doing it
#

I already had docker running and installed, along with Nginx
Jellyfin has better documentation. So I used their guide to reverse proxy Nginx to the Emby service that would come up. All I had to do, was put in my SSL cert paths correctly and change the variable jellyfin to embymusicserver (along with the rest of the require settings.)
I created a folder to hold Emby’s docker compose file. I modified the container name to embymusicserver and gave the path to mount for its config and another one that pointed to my music folder. Pretty straightforward. (I had already changed the ports earlier on the video emby instance to 8097 & 8921 to avoid conflicts with Jellyfin, so I did not have to touch that.)
Assigned ownership recursively to the user that would run the container.
Did a docker compose up to see if everything came up ok. And then tore it down
And finally, once again, adapted my old Jellyfin systemd script to ensure I could have it start and stop with the system

In hindsight, there was no way to pull all of this off in an hour, if it wasn’t for Docker and Emby’s image. Now I realise why containers are taking over the world.

And thank you for all you do, Jellyfin. I shall return as soon as the app gets more useful.

Assigning Keyboard Shortcuts to Nemo Actions

feedback@janusworx.com (Mario Jason Braganza) — Mon, 03 Jun 2024 17:08:35 +0530

Everytime I download a Humble Bundle, I’m left with a ton of books, in various formats on my desktop.
I normally like all the formats for a single title to be in a single folder, so that Calibre can then treat them as a single book¹, when adding them.

I’ve been using Pedro Santana’s, nifty little Nemo action to help me do just this.
Once installed, all I need to do is select my files, right click on them and then create a new folder, like so …

The one remaining wrinkle, being that when you’re completely keyboard driven, you still need to come down a couple of steps with the arrow keys or switch to the mouse.
When doing this with close to 250+ files, this quickly gets tedious.
Stack Overflow to the rescue.

Turns out that all I needed to do was add an _ in front of the letter, I wanted to use to activate my little action from the context menu.
My original action went …

[Nemo Action]

Active=true

Name=Create new folder with selection 
...

I decided to use the n in selection. N for new folder. Easy to remember! So with this tiny change …

Name=Create new folder with selectio_n 

I get to do this!

I can select, get the context menu, hit n, type in the folder name and done!
It’s still a lot of work when there’s lots of files, but there’s a seamless flow to the entire process.

point two, in this section of the documentation ↩︎

Notes to Self on Renewing Wildcard Letsencrypt Certs Manually With Lego

feedback@janusworx.com (Mario Jason Braganza) — Fri, 31 May 2024 11:12:17 +0530

image courtesy, the Lego Github page

I want to someday, automate all my domain certificate renewals.
Namecheap doesn’t support it, ergo my move to Porkbun.
And certbot doesn’t support Porkbun, so I’ve moved to Lego.

I still need to someday figure out how to do this automatically, but for now, since all is in place, this is the new manual process.

You’ll need to generate a Porkbun API key along with its associated secret, and store it somewhere safe and secure. (This is a one time thing, the Generating API keys section here describes how)
Another one time step is to get Lego installed and have it generate your certs, and then have your applications, point to where they are.
Command to renew: PORKBUN_SECRET_API_KEY="your-api-*secret*" PORKBUN_API_KEY="your-porkbun-api-key" ./lego --domains your-domain.tld --domains *.domain.tld --email your-email@domain --dns porkbun renew
And the renewal’s done!
Restart yer engines … err your applications¹ et voilà!
If you have multiple machines, figure out a way to securely transfer the certs there too.

in my case Nginx ↩︎

Using Hugo Variables to Help With Mailto Links in Hugo

feedback@janusworx.com (Mario Jason Braganza) — Thu, 30 May 2024 18:17:35 +0530

The feedback email address seems to be getting traction and I’m getting more mail every week.

I wanted to make it easier on myself and for the kind folks who write in, by including the title of the post in the subject of the email.
Right now, every one tells me what post they are talking about, which as I said in exceedingly kind of the folks who write in. My posts seem to be reaching really considerate people :)

So a teensy bit of head-scratching later, I remembered I use shortcodes to inject custom HTML into my pages. I remembered Hugo has variables I could access. And I remembered mailto links let me add a subject parameter to the link.

So, 1+2+3=Sucess! (Or so I hoped)

The How
#

I looked up the variable, I needed to get at the post’s title. It’s called .Page.Title
I created a hpt.html in my Hugo’s layouts/shortcodes directory to hold the little snippet that’d get typed in the email’s subject field. This is all it contains:
Feedback on post: {{ .Page.Title }}
Hugo will grab this snippet whenever it runs and put in the post’s title in place of the placeholder.
And then I can use that in a mailto link like so, [feedback at this domain](dummyemail@dummydomain.local?subject="{{< hpt >}}")¹

And tada! 1+2+3 does indeed equal success! You can see in action, by clicking on the link below to send me feedback 🙂

This is a markdown formatted link. You can ofcourse use HTML ↩︎

My Experience With the Kinesis Advantage 360 After a Day of Use

feedback@janusworx.com (Mario Jason Braganza) — Wed, 29 May 2024 20:02:51 +0530

Updates : I’ll update this post periodically as I get more familiar with the keyboard

I find the curvy, ungainly shape really beautiful. Click to enlarge

If you follow me on the fediverse, you’ve already read most of this.

Long story short, my wrists (and now, my elbows and arms) are killing me.
So, I saved up and got me a new, split, ergonomic keyboard: The Kinesis Advantage 360
It’s available in the US at around 500USD. I paid close to twice that to import it home, to India (via a seller on the Amazon India site)
Lots of bullet points after a day of use.

All of its ergonomic and software power is QWERTY first and Windows/Mac first. A Dvorak, Linux user like me, will need to adjust quite a bit¹
Everything’s slowed to a crawl as I get used to the new key placements.
The Dvorak layout seems to work better on this than on my old keyboard
The key layout and spacing is throwing everything off
Realised some of my keyboard shortcuts were that way, because of the old keyboard layout. They are terrible on the new keyboard. Need to find new shortcuts for ulauncher, minimising windows, and the gpaste clipboard manager
My caps lock is the compose key and i need to see if the new position will help or hinder me
totally gotten used to the new space, return and backspace keys
totally not used to the placement of esc. will try a week to see if i can get used to it.
new ways of chording keys in emacs. seem to be easier on the wrists. (because of the new place for the ctrl and alt keys)
my wrists feel much better. no idea how much of this is placebo. will know for sure only after a while
bottom row keys are tough right now
got used to the compose key
the question mark and slash has moved below and to the right (dvorak), so will need to see if i adjust
i miss my numpad so much!
learned how to mount the onboard flash storage and eject it properly
updated firmware to the latest version.
also note to self, switch to querty before you do this. status reports are typed out wonky other wise
I have brown switches

The one thing I confidently say, even after a single day of use, is that this is a lovely keyboard.
I will get used to all its quirks, because the scalloped, split design suits me really well.
Most of it, is just changing muscle memory, not learning new painful maneuvers.
And the design also ensures, every single key is in reach of my fingers. This is the magical part for me.
If it lasts me five years, I’ll consider it money well spent.

Updates
#

2024-07-02
#

Gotten used to the layout for the most part. The arrow keys are still very troublesome and trip me up all the time. Realise how deep muscle memory actually is.
The only reason I don’t want to change them, is that they (the new keys) are easy on my fingers and when I’m intentional about them, are easy to use.
The Compose key and the : keys will still take a while getting used to.
Have gotten the hang of /, ?, = and + though.
Have gotten back to my old speed. My error rate is still really high. Most of my friends have now gotten used to my typos 😂

for e.g., the [ key finds itself at the top right, while the ] key finds itself top left 😂 ↩︎

Ignoring Syncthing Conflict Files in the Org Agenda Folder

feedback@janusworx.com (Mario Jason Braganza) — Sun, 26 May 2024 16:03:01 +0530

Now that I have the power of large language models, in the palm of my hand …

You just had to be there, to really appreciate this 😂
Image, courtesy user straiven on makeagif

I decided to see, if it could help me solve one of my real pain points with Org Mode.
I had asked for help over on the fediverse, after months of trial and error, on just how to fix this.
The issue being, I needed a way to tell my Org Agenda, to ignore files with a certain pattern.

I have a folder (called syncbox) full of files, that Org Mode looks at to generate the agenda for me.
Now these files are also synced via Syncthing over to my phone, where another program called Beorg looks at them too and shows them as entries in my calendar, so that I don’t miss the really important things I need to do.

Functionally, all is well with this setup, except for one teensy thing.
Every once in a while, when certain files are in use on both, my desktop and my phone, Syncthing will create conflicting copies of said files in use.
This happens most often, with my tasks.org file.
Once it conflicts, my syncbox gets full with files like,
tasks.sync-conflict-20240408-041556-2TQWENY.org and
tasks.sync-conflict-20240516-101335-2TQWENY.org and
tasks.sync-conflict-20240520-053451-2TQWENY.org

What happens then, is that Org Agenda happily reads those files too, and I end up with lots of duplicated tasks.¹

As usual, tackling it by writing lisp, was an idea fraught with terror for me. I tried quite a bit and failed.
My next plan of attack was that Org Agenda had a regex², that it would process. My regex is better than my lisp, but even that was to no avail.
Emacs regex does not support lookaheads / lookbehinds and I did not want to have some complicated pattern, that even I would not recognise, a few months down the line.
In the meanwhile, Ihor told me to stop faffing about and just write a function³

So I asked Claude to go do the work for me and I ended up with …

(defun mjb/set-org-agenda-files-from-syncbox-directory ()
 "Set `org-agenda-files' to all Org files in the 'syncbox' directory, excluding files with '.sync-conflict' in the name."
 (let ((syncbox-dir (expand-file-name "syncbox" my-org-directory)))
 (when (file-directory-p syncbox-dir)
 (setq org-agenda-files
 (mapcar
 (lambda (file) (expand-file-name file syncbox-dir))
 (seq-filter
 (lambda (file) (not (string-match-p "\\.sync-conflict" file)))
 (directory-files syncbox-dir t "\.org$")))))))

(mjb/set-org-agenda-files-from-syncbox-directory)

which I dutifully dumped into my init file.
And that did the trick!

What did strike me, is that now that I am using this approach, my chicken and egg struggle is finally at an end. I don’t have to struggle to learn lisp to hack emacs which needs me to learn lisp.
I realise it takes only a short time of dedicated effort. But right now, it is time I don’t have, and don’t know when I’ll get.
So the new fangled LLM AIs can help me sort my Emacs pain points and I can learn Lisp at leisure.
And I’ll use, friend on the fediverse, Vincent Dardel’s excellent course to do just that.

It drove me nuts in the beginning, when I did not know what was happening. I was like, I /just/ finished this task. And it’s popping up again?! Is my computer haunted? ↩︎
defined in org-agenda-file-regexp with a default value of "\\`[^.].*\\.org\\'" ↩︎
He did not quite put it that way. He was much kinder, obviously :) ↩︎

Org Mode Clocking in and Out of Tasks on Task State Changes

feedback@janusworx.com (Mario Jason Braganza) — Thu, 23 May 2024 19:03:13 +0530

Logo courtesy, the Org Mode Website

Lisp is on the list of things to learn someday.¹
And in the meanwhile, I need to be able to hack Emacs to do the stuff I want.
But also, there are a ton of other things, that are currently much, much higher on my list of priorities.²
And while the community is extremely kind and generous, the expectation is that you need to put in the time and the work.
Which I currently don’t have and can’t do.

And so, it finally happened.
I signed up for a Claude account and told it to write me some lisp.

I am heavily Org Mode task driven, these days and I have a manual workflow that I need a teensy bit of help with.

I start work on a task
I mark it as WORKING (and clock in manually)
I want to switch to another task
I mark my original task as WAITING (and clock out manually)
I then mark my new task as WORKING (and clock in manually)
After a while I switch to the original task, and clock in manually (after marking the current task as WAITING and clocking out there manually)

What bugs me, is the clocking in and out, because …

I tend to forget it, messing up all my clock times and …
because it is now becoming painful on my middle aged fingers³,

So, like I said, I prompted Claude to give me code that would clock me in automatically, whenever I changed state to WORKING and clock me out when I switched to WAITING⁴
It spat out some code.
I tried to reasonably guess and verify what it did.
It looked ok.
I tried it.
It worked!
And that’s that.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17


(defun mjb/org-clock-in-when-task-state-changes-to-working ()
 "Clock in an Org mode task when the status is changed to 'WORKING'."
 (when (and (eq major-mode 'org-mode)
 (member (downcase (or (org-get-todo-state) "")) '("working")))
 (org-clock-in)))


(defun mjb/org-clock-out-when-task-state-changes-to-waiting ()
 "Clock out an Org mode task when the status is changed to 'WAITING'."
 (when (and (eq major-mode 'org-mode)
 (member (downcase (or (org-get-todo-state) "")) '("waiting")))
 (org-clock-out)))

;; These lines make the functions above, hook to Org’s state change mechanism.
;; So that everytime a task’s state changes, these functions are called
(add-hook 'org-after-todo-state-change-hook #'mjb/org-clock-in-when-task-state-changes-to-working)
(add-hook 'org-after-todo-state-change-hook #'mjb/org-clock-out-when-task-state-changes-to-waiting)

Reminds me of the old adages about humans and bicycles being the fastest and humans and chess programs⁵, being the best players.⁶
All of it is good, if it aids me, and hopefully does not get in my way.

soon ↩︎
getting up to speed with the current devops/cloud landscape, for instance ↩︎
Yes, I could reassign them to easier keys, but just having it done automatically is better ↩︎
the task state switches are second nature and don’t bother me much ↩︎
Centaurs, I think, the combination is called ↩︎
and of course, the pilots and giant mechas in Pacific Rim ↩︎

Rust

feedback@janusworx.com (Mario Jason Braganza) — Wed, 22 May 2024 16:32:40 +0530

Alright!
Trying to get back to writing. (attempt no. 87,245)

And trying to learn Rust and document the journey at the same time.

Why?
#

I want to learn a language that can give me free standing executables to distribute.¹
I want a fast language. (rather a language that generates fast programs.)
The little that I’ve learnt of C (which fascinates me), has convinced me that I am one of those folks, who will definitely shoot myself in the foot with it. And most stuff is manual. I’m willing to do the work, but not all the work. Lazy and all that :)
That boils it down to C++ & Rust.
To my mind, it’s an easier path from Python to Rust. The shapes of things are similar. Like Fred Ross would put it, it’s not an ur-language for me.²
And at this point, it was just looking at a Rust error message that did it for me. Short. Sane. Sensible.³

The Plan
#

There’s no grand plan here, unlike Python where I just had to know everything
Just do a tiny bit daily and write about it. That’s it.

It’s hard with Python. ↩︎
Lisp is! :) ↩︎
I hope the documentation is as good. I haven’t read much. The little that I have befuddled me. But that is just because of the new, new thing. ↩︎

Note to Self Footnotes Using Littlefoot.js

feedback@janusworx.com (Mario Jason Braganza) — Wed, 28 Feb 2024 08:10:36 +0530

Courtesy dothegreenthing, CC BY-NC 2.0, on Flickr

I have this amazing urge never to leave well enough alone.¹
The blog uses an amazing theme, called PaperMod.
And ofcourse, like I said, because I can’t leave well enough alone, I’ve customised it heavily.²
The fonts are different, there are various kinds of slight colour changes, I bullied Aditya into showing me how to build a Menu, there are lots of custom Hugo shortcodes and there were footnotes powered by a nifty little Javascript library called littlefoot.js.

Now everything was fine and dandy.
There are folks who do leave well enough alone. Their blogs stay up and running for years on end, held together with nothing more than spit and baling wire.
Those folks aren’t me, sadly. Aditya keeps improving the theme and I really wanted the new shiny. So I went and got it.
And everything broke!
I got weird artifacts. RSS broke. Fonts broke. Footnotes broke.³

I’ve managed to put everything back together, over the past week, and I just got footnotes working again yesterday.
So a few notes to myself, on what to do, when I break it again.⁴
Footnotes today, and the rest later …

So here goes,

The litlefoot.js repo has excellent documentation. That’s where you pulled everything from (for when you wonder, just how you figured all the CSS out)
You download a static copy with a npm install --save littlefoot and then stick it into Hugo’s static folder
Copy the extend_head.html from the theme’s layout/partials into your own. You should already have an extend_footer.html since you put in PaperMod’s hamburger stuff in there.
Adapt the Usage from a CDN step to directly access it from /littlefoot/dist/… instead of UNPKG. Put the CSS link into your extend_head.html and the JS into extend_footer.html
We want the footnotes to be numbers, rather than floating ellipses. So modify the default (blank) script call in extend_footer.html, to the stuff below. This is verbatim from the docs.

<script type="application/javascript">
 littlefoot.littlefoot(
 {
 buttonTemplate: `
 <button
 aria-label="Footnote <% number %>"
 class="littlefoot__button"
 id="<% reference %>"
 title="See Footnote <% number %>"
 />
 <% number %>
 </button>
 `
 }
 )
</script>

We want to style it a teensy bit, so those styles go into extend_head.html. For some reason, it does not pick the footnote styling correctly when you put stuff into your normal assets/css/extended/ folder.

<!--Custom littlefoot styles-->
<style>
 .littlefoot {
		/*the --attributes come from the littlefoot.js docs*/
		/*Change button background to something that works in both modes*/
 --button-background-color: #666; 
		/*Change padding to make the button rounder*/
 --button-padding: 0 0.35rem;
		/*The footnote ought to be italicised*/
 --popover-font-style: italic;
		/*Make the popover work with my theme’s colour scheme*/
 --popover-background-color: var(--theme);
 --popover-text-color: var(--primary);
 /*the oldstyle numerals used everywhere else, don’t look good here
		Neither do they align properly in the box.
		So switch them to normal*/
		font-variant-numeric: normal;
		/*Italic numerals for the footnotes*/
 font-style: italic;
 }

 .littlefoot__button {
		/*very subjective. 
		try to get the numerals centred as much as I can.*/
 line-height: 1;
 vertical-align: middle!important;
 }
</style>

Et voilà! Footnotes galore!

It looks a little weird, a little like a billie goat walked across my post, pooping pellets across the page.⁵
But it is my kind of weird now.
And I love it :)

what we in the Mumbaiyya vulgar, would call, “Gaand mein keeda”.
Please for the love of God, don’t look that up. Very not safe. ↩︎
Heavily, as in heavily for a guy who doesn’t quite know what he’s doing. ↩︎
All my fault obviously, with all the custom stuff. Nothing to do with the theme proper. ↩︎
in the far, far, far, future; future Jason! Not any time soon, you hear! ↩︎
“Pooping pellets on the posted page”, Now there’s a tongue twister 😂 ↩︎

Org Mode Narrowing and Widening Buffers

feedback@janusworx.com (Mario Jason Braganza) — Mon, 26 Feb 2024 09:40:17 +0530

tldr: C-x n s to narrow, C-x n w to widen

Now that life’s getting busier and busier, with lots more things on my plate, ye olde system of writing the day’s tasks down on paper does not work any more. Specially since, I’ve begun logging clock times¹. Writing them down by hand is a giant pain.
This is what the plan’s begun to look like.

An unwieldy plan.
Click this image or any below, for larger versions

I wanted to see, if Org would let me replicate what I did my hand, natively.
Which in a nutshell was to look at the day sheet (on paper) for the nitty gritty and my organiser (currently my tasks.org file) for the big picture view.
I bounce between the two often. And discrete focus is something I wanted to keep.

The focus from wide …

to narrow. And back …

A bit of searching and this is the result! Woohoo!

Like I mentioned above, I hit C-x n s next to my day’s headline to narrow the scope to just today. And a C-x n w gets me back to the thirty thousand foot view.
I think of them as, s for slim and w for wide, to keep them straight in my head.
I’m intuiting this is Org’s Sparse Tree Magic at play here. But learning that wizardry is a thing for another day.

The important bit for me, is that I can focus when I working, without getting distracted.
And then when I need to, zoom out to review what progress looks like and make changes, if I need to.

I feel a bit sad though. I’ve given up book notes on paper. I’m giving up planning my day on paper. It feels like I’m losing something tangible. I miss paper.

I hope to use these, to objectively measure where my time is really going ↩︎

Note to Self, Bookmark Your Work, Coda

feedback@janusworx.com (Mario Jason Braganza) — Sun, 25 Feb 2024 10:24:45 +0530

courtesy Liniers, Comics Kingdom
click to enlarge

With all the hullabaloo yesterday, on how you’d go about organising your day, and things to let go of, and marking your place, I forgot the overarching Note To Self.
Please remember. Never lose sight of what’s important.
Never forget why you’re doing all of this.

Note to Self, Bookmark Your Work

feedback@janusworx.com (Mario Jason Braganza) — Sat, 24 Feb 2024 08:34:05 +0530

This started with a note to self on leaving breadcrumbs, and ended with a stream of consciousness journey on how I reached this conclusion. So what I’m trying to say, in a slightly rambling manner, is that this is a slightly incoherent, rambling post.

Courtesy, Delgrosso, CC BY-NC-ND 2.0, on Flickr

I’m running my days now to slow, daily, consistent tiny amounts of progress.
And I am seeing results. When there are so many demands on my time, prioritising them and then doing them slowly yet consistently is yielding fruit.
While I certainly enjoy the deep dive, full of focus, lost in work sessions a lot more, I have too much life happening to allocate that kind of time and attention to everything.
The first thing I did, when I kinda-sorta revamped the way I went about my days¹, was to drop as many hot potatoes as I could, and then d-d-d-ed all the small yet important things in my life.
And then came to the sobering realisation that the rest of life just had to be done and dealt with. Even things I don’t like.

A couple of pieces then had to fall into place, before I gained some semblance of sanity and progress.

I can’t do it all. Some things, important things even, will fall through the cracks. Somethings that I really want to do, will get pushed to someday/maybe and some will just never get done. I have to make my peace with it.
Knowing what I am giving up, the things that I am doing? They better be worth it.
Making the time to think this through, and having mental muscle to do the thinking² helped.
Everything does not have to be good and perfect. Good enough, is good enough.
Yes, stuff is hard and unpleasant, but then you have a blessed life. There are so many people with hard lives. Suck it up. And go do something about it. Deal with it. Leaving stuff unattended in the recesses of your mind, is a sure way to rust and ruin.³. At the same time, while you are hard and focussed on the work, be kind to yourself. Find that balance.
Alright, mentally set. Now how do I go about the doing?
By doing as much or as little as you can, depending on how you prioritise it.
By doing it consistently. Set a schedule. Hit it.
If you think you don’t want to do it, do 10m regardless.⁴ And then move on.⁵
If interrupted, deal with the interruption. Come back. And continue.
Dealing with interruptions does not mean giving in to it. It means figuring out just what it is, and figuring out what to do about it. Most stuff can be jotted down and scheduled for later. Make sure the significant folk in your life know what you are up to, so that they know when they can expect you to be present and when you are off doing your own thing.⁶
Doing stuff this way is slow. And that’s ok. If your life permitted focussed deep work, you’d be doing it. But that’s not how it is. So find islands of focus. And do the best you can.
Do the best you can, does not mean going all out. Slow, sure, steady. That is your drumbeat. Slow, sure, steady. Unflustered. Calm. Measured.

So what happens once you do this? And what hiccoughs did you run into? And what do you do about it?

Stuff gets done! It doesn’t look like this ought to work, but it does.
Leaving stuff undone, putting things down for the day, causes lots of mental turmoil.
Also whenever you go back to stuff, there’s a lot of going back before your minds grasps where it was at, and only then does it go forth.
And that brings me to the crux of why I began writing this dastardly note in the first place.
Log. Summarise. Review!
1. Log how long you take for your tasks. Don’t judge yourself. Just dispassionately do it.
2. Every time you think you’re done with something for the time being and are putting it down? Spend the time to write down what you did and what future you needs to be doing. Create bookmarks. Give directions. Do this no matter how tired you are. No matter how shitty the note.
3. Review. This is the most important bit. Review at regular intervals and at various depths.⁷ And only then make your plans accordingly. Daily time logging & summarising leads to daily review leads to a better sense of what is done, how far you’ve come, what to change, where to swerve, what to do next, and where to go. Review, gosh darn it. Review!

It’s been working well the past three months. It’s reduced a lot of friction, and quite a bit of time. It makes it easy to get into the swing of things.
I’ll come back in six months and write about how it went.

it was a slow, gradual process. ↩︎
Thank you, meditation practice! ↩︎
oft used phrase in the Cosmere ↩︎
be as foul minded and potty mouthed and grumpy as you want. But do it. ↩︎
unless you get into the groove. ↩︎
if you are busy doing work all day, then that’s a sign. You need to get your life back into some semblance of balance. It’s ok doing this shit for a measure of time. This cannot be your whole life ↩︎
I’m good at the nitty gritty and the big picture (daily / quarterly reviews). I struggle with the medium term (weekly) ↩︎

Note to Self: Less Research, Less Rabbit Holes

feedback@janusworx.com (Mario Jason Braganza) — Sat, 17 Feb 2024 09:56:11 +0530

Alice’s Rabbit Hole, Sandra Strait / CC BY-NC-ND 2.0
Click on the image for the larger original.

Big realisation, I am procrastinating on my other, bigger priorities, with the 100WordHabit project.
Bigger realisation, if not this, I’d find something else.
Biggest revelation, even if life sucks donkey balls, I need to focus, be disciplined and do what needs doing.

So here’s a few stream-of-consciousness observations & suggestions, to future Jason, when doing the project.

The writing almost never takes you that long. You write at a fairly decent clip.
What does take you a lot of time, is that while you do know what to say, you take an inordinately long time, trying to figure out just the right way to say it, to present it.
It’s the rabbit holes that you get into.
a. You take too long researching, finding just the right image and the right words.
b. You just spent half an hour trying to figure out the right way, to attribute an image.
c. You spent another trying to find out how to jump to a specific column in Emacs (M-g followed by tab followed by the column number and return)
d. You started curating and tagging some mp3s that have being lying around for years now.
e. then you went and forked the music jukebox library you used on Github
You hate doing the hard stuff.
a. It makes you really uncomfortable
b. It makes you want to take breaks
c. You keep switching to obssessively checking all your feeds.¹
d. You then find something that you just have to learn more about and that leads you into another rabbit hole.
e. Besides you’re a completionist. All those tabs and notes to self, just have to be read and squirreled away
All of this is taking time away from stuff you really have to do considering your age.
a. You hardly get time to exercise. Yes life is busy, but you don’t have to make it worse.
b. The career pivot is going slower due to this.
c. You don’t have time to execute your diet.

Here’s a few things you could do …

Remember, this is only during your deep work hours. We need to make the most of them. Go nuts when you aren’t working.²
Figure out what your process is. I realise you’re a pantser, but a little bit of outlining will do you no harm. You already have a running list of topics to prevent writer’s block and time waste. You’ve already lowered the bar, to become more consistent. So see what else, you do in your writing process that takes the longest. In fact, what do you do? Will recording yourself help?
(Added, 2024-02-18,) Don’t check your feeds in the mornings. This one’s a huge timesink. You find issues that interest you. Then you go to find someone who speaks with a nuanced voice. Nuanced voices are rare. They’re hard to find! And if you decide to listen to Munger, and go do the work to have an opinion, by trying to find an opposing nuanced voice? That’ll take you even longer. Just STOP! Now is not the time to get better informed. Do it later.
Start clocking times with Org. You want to get a better idea of where the time goes.
Don’t spend too long reading the manual. Just figure out how to clock in, clock out, get a summary
Set a time block? Posts just have to be published once you hit a hundred words or a hour, whichever is longer? If you still want to continue, do it the next day. The posts could be part one / two / three / blah?
IT DOES NOT HAVE TO BE PERFECT! Be ok with “mediocre”, unstructured posts. This phase will not last long.
Don’t let this take over your life. You already have one vice.

The most important thing though? Kudos to you!
The meditation practice is really paying off in spades. You notice things. And you try and do something about it. You don’t let it bother your present. Most importantly, you don’t let it devolve into self loathing. You’re a lot less stressed, a lot more peaceful and a lot more resilient than you used to be.

courtesy Liniers / Comics Kingdom

There’s nothing new there, since the last time you checked, you twit! ↩︎
The point is also to make the time ↩︎

The Kobo Elipsa 2e, Six Months Later

feedback@janusworx.com (Mario Jason Braganza) — Fri, 16 Feb 2024 09:28:55 +0530

My beloved, battered, Elipsa 2e. Click the image for a larger image.

It’s been close to ~~six~~ seven months, since I moved over from the Oasis to the Elipsa.

Here’s my list of things / experiences with the Elipsa 2e, positive and negative, in no order whatsever. Also this is not representative of a typical ebook reader¹. I, umm … have a lot of books, and the Elipsa carries most of them. So every action that it does, it does with that number of books (either physically or with its database) …

I love it!
It’s very light for its size. I don’t feel the weight of this 10″ device as much as I felt the old 7″ Oasis.
It’s very fast.
I use it just like a book. It lends itself to such use as well. Part of it is the weight, part of it, is its responsiveness.
This is also how I broke it. I threw it on the bed and then it slipped right over the edge, and cracked the screen 😂 You can see the big chonky part that is permanantly dead, as well as the thin, dead, vertical and horizontal lines. Since I read with fat margins, the chonky part does not get in my way² and the lines are mostly invisible.
~~I cannot enlarge images in a book. I tap, I double tap, I pinch to zoom … nothing. This is the one thing I miss from the Oasis.~~ A few hours later: Of course, now it works! 😂 Double tapping an image, opens up the image where I can zoom and pan around! My fault for not noticing this earlier.
I love it!
It hangs loading complicated books or large pdfs, but it recovers quickly. This does not happen that often.
It randomly hangs and/or crashes. This happens quite often. But I’ve learnt to reboot it with the power button, and it comes right back up, quickly. It reminds me of the Am386 machines, I’d assemble for use as Netware thin clients in my youth. They would run hot and crash ofen, but we didn’t mind, because they rebooted right back so darned fast 😂
Not that the Elipsa runs hot. It just runs fast and crashes. And then comes back on and carries on as if nothing ever happened.
Did I tell you, I love it?
On paper, the screen specs look inferior to the Oasis. In practice, it made no difference to these old eyes.
The user experience is not as good is it could be, but it’s miles better than anything else I’ve used. I figured out most of it intuitively. I love just running my finger along the left edge to adjust the brightness. Helps a lot when I’m reading in bed, so that the better half does not get disturbed (or grumble at me much.)
Books on the Kobo store are a bit more expensive than the Amazon store. (New releases tend to cost the same.)
The database that it uses, gets corrupted quite often, which then leads to it no longer being able to talk to my PC or not showing the books in the device³. I doubt, I’d love the Elipsa as much as I do, if the Calibre ecosystem didn’t support it so well.⁴
- I can backup and restore the database at will, thanks to the awesome Kobo Utilities plugin.
- I’m getting old and impatient and want to read a lot more. So I’ve given up on taking notes by hand and moved over to highlighting and making notes in the Elipsa. I wouldn’t do this if not for the lovely Annotations plugin, which lets me get all my notes and highlights out of the Elipsa and into Calibre. (pic below, click to embiggen) The jury’s still out on this one.
I love it so much, that when I heard a friend was going to the USA, I bullied them into buying me another one to replace this one when it dies.

To conclude, the Elipsa 2e is wonderful.
It has flaws that I can live with.
It serves me well.
I love it!

the person, not the device ↩︎
much ↩︎
This point I think, is more a me problem, than a device problem ↩︎
Note to the Kobo folk, please care about your device software a smidgen. Please! At least at as much, if not more than these lovely open source folk do! ↩︎

Note to Self, Emacs Help Stuff

feedback@janusworx.com (Mario Jason Braganza) — Thu, 15 Feb 2024 08:03:24 +0530

Collection of various little things I find handy when I am lost in Emacs

Truly lost and need help about help itself? C-h C-h or C-h ?
C-h m should bring up a list of all the shortcuts available to that mode.
Want to find out what the docs mean when they refer to something? C-h o and then type in that name.
For e.g. when the Org Mode docs refer to org-hide-leading-stars, a C-h o => Return => org-hide-leading-stars will give me details about what it is. (According to the docs: org-hide-leading-stars is a variable defined in ‘org.el’. Its value is nil.) Alternatively if you know the type of object, you’re looking at, C-h f will give you help on a function and C-h v will do the same with a variable.
If you want to know what function is behind this crazy shortcut you’re typing, a C-h k followed by giving it the shortcut your typing will lead you to the ~~culprit~~ command that powers it.
For e.g. I blanked out on how to toggle multiple checkboxes in Org Mode today and I was hitting C-c C-x b to on avail. Running C-h k told me that it was bound to org-tree-to-indirect-buffer. So I was madly mashing the wrong keys. A little digging (with C-h m) revealed that what I needed was C-c C-x C-b which ran org-toggle-checkbox which indeed did what I wanted! :)

My Raspberry Pi 5 Ran Hot

feedback@janusworx.com (Mario Jason Braganza) — Tue, 13 Feb 2024 13:52:09 +0530

I got my greedy paws on a Raspberry Pi 5 last month.
Probably the Pi after this, will be the one at which I will say these are good enough and stop my “get ’em as soon as they are available” manic buying spree over the past few years.

I use it as a music jukebox (with Jellyfin), as a place to host my audiobooks/files, as well as a syncthing node. Occasionally I will download and transcode videos I find interesting, with yt-dlp.

The Pi is blazingly fast at these tasks. Stuff gets done in a third to half the time, they would take on the old Pi.
It also gets blazingly hot apparently.
Which reminds me of my time assembling 486 desktops in the late 90s. The DX2s needed only heatsinks (like the Pi 4) whilst the DX4 needed fans as well (like the Pi 5)
Like they say, everything old is new again.
It also reminds me of the time the fans would just up and die, after which the machine would then heat up and just go off.
The exact same thing is happening with my Pi 5. The fan in the official case is close to useless. It started whining loudly after only a week of use. But even when it did come on, the Pi would heat up and die.
I thought this was because of the power adapter I was using. Official chargers are not available here in India, for love or money. Which is why I was using the highest rated mobile charger I could lay my hands on. To eliminate this as a cause, I switched to a 65W Anker adapter. No dice.
I installed the sensors package, to get an idea of what was happening. The cores were spiking to 80°C plus. No wonder!
Went looking for heatsinks or fans that could assuage this issue and found an all metal Verilux Aluminium Alloy Case. Though calling it a case would be a slight misnomer. What it is, is sheets of aluminium, separated by spacers, topped with a fan.

It seemed overkill, but at the time, I had no other option available.
I got it, screwed everthing in, praying all the time, that I wouldn’t lose a spacer or screw. (They give one extra screw, nut and spacer)
That did the trick. Temperatures are in the high 40s to low 50s. And everything is zipping along and hunky dory again.

Get Hugo to Show Images in Links When Shared on Social Networks

feedback@janusworx.com (Mario Jason Braganza) — Sun, 11 Feb 2024 19:45:52 +0530

Whenever I post a link to one of my posts on the fediverse or share it with my friends on Signal, I just get a plain vanilla link box

I wanted those rich embeds, other folk seemed to have, on the same mediums, that I would share my links on.
Looking around for how to go about doing this, led me to ogp.me, which taught me the whys and wherefores.
Looking around, a bit more for how to do this with Hugo, led me to this thread, which had this helpful comment by Kaushal Modi, which made be search the Hugo documentation which finally led me here.

I figured all I needed to do then was just to link to the image¹ I wanted in the metadata section of my post, like I did for this one.²

title: "Get Hugo to Show Images in Links When Shared on Social Networks"
date: 2024-02-11T19:45:52+05:30
categories: ["work"]
tags: ["hugo"]
images: 
 - "/images/2024/op-image-on-link-2.png"

Et voilà! Rich image previews on links!

or a list of images, if you so choose ↩︎
I use yaml. The docs have examples for toml and json formats as well ↩︎

Note to Self, Switch to Manjaro “Unstable”

feedback@janusworx.com (Mario Jason Braganza) — Tue, 06 Feb 2024 09:01:31 +0530

I kept wondering why Syncthing releases on one of my Pis would lag behind my other one.¹ The big difference between the two is that one runs Raspbian and the other, Manjaro²³. This wouldn’t happen earlier when they ran Arch and Manjaro.
And today, I head-slappingly remembered, that it did lag and I had to do stuff to get at the latest software stuff.

What I had to do, was to switch Manjaro’s software branch to unstable (as opposed to the default stable. Read more here.)
The only reason I use Manjaro, instead of Arch⁴, is that the Arch kernel does not boot on this Pi and I want the latest Arch stuff, which is what the unstable Manjaro software branch offers.

So a …

sudo pacman-mirrors --api --set-branch unstable

switches me to the unstable branch and …

sudo pacman-mirrors --fasttrack 5 && sudo pacman -Syu 

will resync all the software with what’s available on that branch.

Tada, I have syncthing v1.27.2 "Gold Grasshopper" on all my machines :)

as well as other software in general ↩︎
because, reasons. and a long chain of upgrade and experimentation and moves from Pis 3 to 4 to 5 ↩︎
It’s the Manjaro Pi that lags. And yes, I know you can already guess why ↩︎
with apologies to Manjaro ↩︎

Work on Janusworx

Moved From Papermod to Blowfish

Why move to Blowfish? #

Blowfish specific things for future Jason to note. #

Big Picture things #

Configuration #

Content Structure #

Archetypes #

i18n (Internationalisation) #

Layout Overrides #

RSS #

Search #

Littlefoot #

Redirect #

Shortcodes #

Static + Tools #

Themes #

CSS #

Miscellanea #

Go Templates: #

You Can Now Search this Blog

tldr: https://janusworx.com/search #

The Plan #

Installing it and wiring it all up #

1. Installation #

2. Creating an entry for search in the navigation menu #

3. Creating the search page. #

a. I needed a placeholder page that I could render some layout (a template) into. #

b. Creating the search layout. #

4. Make the page resemble the rest of the blog as much as possible. #

5. Make Search Work. #

a. Build the blog. #

b. Get Pagefind to create its index in the build directory. #

c. Publish! #

Matthias Endler Built His Own Newsletter Setup

Found MJML

Kubernetes & CNCF Awards

Trying a Different Tack to Writing Work Posts

The Big Plan: Change My VM to Be Gitops Driven

List of services that absolutely need to come over. Miscellaneous stuff later. #

The Big Point of The Big Plan #

The Big Plan (Done! 🎉🎉🎉) #

Unrelated. Long term. Optional. Just here so that I remember #

Moved From Reeder to Readkit

Kubernetes, Note to Self: Need Load Balancer Installed on Bare Metal

K3s: Move Data Folder

Problem #

Solution #

Hugo Redirect From Post in One Section to Another

Problem #

Solution #

The Plan for My New Hetzner VM

Update: 2025-12-11 #

Preamble #

Goals #

What I want #

2025-11-06 09:35 #

2025-11-06 16:00 #

Make a Ruckus: 001, Done! Or Are We? Thank you, Dave!

Worth it financially? #

Worth it for the knowledge I gained? #

Am I done with this series? #

Dave as teacher and all round good egg #

Make a Ruckus: 000, The Motivation

Why? #

Emacs Package Updation Checklist

Forgejo Workflow Action to Publish Hugo Blog

Here’s What is Stopping Us From Using Free Software

Blog Questions Challenge 2025

1. Why did you make the blog in the first place? #

2. What platform are you using to manage your blog and why did you choose it? #

3. Have you blogged on other platforms before? #

4. How do you write your posts? #

5. When do you feel most inspired to write? #

6. Do you publish immediately after writing or do you let it simmer a bit as a draft? #

7. Your favorite post on your blog? #

8. Any future plans for your blog? Maybe a redesign, changing the tag system, etc.? #

Hack Hugo Post Metadata With Python

Tmux Start Session Maximized With Three Panes

View Only the Day, With Org Agenda

Why move to Blowfish?
#

Blowfish specific things for future Jason to note.
#

Big Picture things
#

Configuration
#

Content Structure
#

Archetypes
#

i18n (Internationalisation)
#

Layout Overrides
#

RSS
#

Search
#

Littlefoot
#

Redirect
#

Shortcodes
#

Static + Tools
#

Themes
#

CSS
#

Miscellanea
#

Go Templates:
#

tldr: https://janusworx.com/search
#

The Plan
#

Installing it and wiring it all up
#

1. Installation
#

2. Creating an entry for search in the navigation menu
#

3. Creating the search page.
#

a. I needed a placeholder page that I could render some layout (a template) into.
#

b. Creating the search layout.
#

4. Make the page resemble the rest of the blog as much as possible.
#

5. Make Search Work.
#

a. Build the blog.
#

b. Get Pagefind to create its index in the build directory.
#

c. Publish!
#

List of services that absolutely need to come over. Miscellaneous stuff later.
#

The Big Point of The Big Plan
#

The Big Plan (Done! 🎉🎉🎉)
#

Unrelated. Long term. Optional. Just here so that I remember
#

Problem
#

Solution
#

Problem
#

Solution
#

Update: 2025-12-11
#

Preamble
#

Goals
#

What I want
#

2025-11-06 09:35
#

2025-11-06 16:00
#

Worth it financially?
#

Worth it for the knowledge I gained?
#

Am I done with this series?
#

Dave as teacher and all round good egg
#

Why?
#

1. Why did you make the blog in the first place?
#

2. What platform are you using to manage your blog and why did you choose it?
#

3. Have you blogged on other platforms before?
#

4. How do you write your posts?
#

5. When do you feel most inspired to write?
#

6. Do you publish immediately after writing or do you let it simmer a bit as a draft?
#

7. Your favorite post on your blog?
#

8. Any future plans for your blog? Maybe a redesign, changing the tag system, etc.?
#

Emacs
#

Rich Text Expansions
#

How does this affect you, my kind and gentle reader?
#

Setting the stage
#

Doing it
#

The How
#

Updates
#

2024-07-02
#

Why?
#

The Plan
#