Few hours ago, I uninstalled Instagram from my phone.¹

Ironically, earlier today I watched a reel on Instagram saying “Instagram is like Chakravyuh, and we are like Abhimanyu. We can go inside, but not come out”.²

I think that made to take this action.

I noticed that I had not written much in the entire month of August 2023, and hardly 2 entries in September. While some of it attributes to work pressure, my instagram addiction may be equally responsible.

As mentioned earlier I tried new starter kit called Emacs Bedrock.

Getting started was easy with --init-directory option in Emacs 29

I also enabled a few mixins in init.el as recommended. Especially evil mode.

The *Quick Help* that shows up at the start is very useful. Especially for someone like myself who is not familiar with native emacs keybindings.

Enabling evil mode means I can at least edit the text.

But there are so many thing one does other than editing the text.

Recently when reading about emacs Bedrock I learnt that emacs 29+ has command line option that lets the user pass custom init directory via emacs --init-directory=MYDIR option. ¹

This makes it easy to try out different starter kits without touching your stable/working config

Currently, I use Minemacs on my personal machine. In past I would rename my .emacs.d to try new config. (Then switch it back)

This makes things easier.

I’ve been using vi for 20+ years and somehow I didn’t know that it has spell check built in 🤷‍♂

Enable it by simply :set spell

Now misspelled words are shown in red background ¹ (depending on the terminal’s capability this might vary - I assume)

Off course, this is nowhere near full fledged LSP like ltex which I use with emacs and helix. ltext does more than spell check.

Yet having built-in spell checker is always nice.

My thoughts in the context of this thread.

I did try org for some time (initially when I switched to Emacs) as mentioned here

But eventually I realized that markdown is more pervasive.

I am aware that Org is supported outside of Emacs as well.

But compared to markdown, that is nothing.

But that is not necessarily the point.

e.g. When I do code review, most of the git service providers support markdown. I was surpised that even Azure git supports Markdown (So does gitlab, probably bitbucket as well, it as been a while since I used bitbucket.)

Why do we “listen” to somebody ?

Here I do not mean hearing when I say listen

I mean obey or follow.

I think it comes from either respect or authority.

Authority maybe given by someone else

e.g. Teacher in school, boss at work, parents at home.

Lot of times, most people do what these people tell you to do, because of the authority inferred on them.

Because you are “supposed to” do what they say.

I had written earlier how learning the language before the framework is more important.

It is important to know which parts is part of the language, and which comes from the framework.

This becomes more important when there are defacto web frameworks (rails in ruby, django in python, phoenix in elixir.. maybe more) For a newcomer who is learning both the language and framework together (or back to back, or worse learning the language via the framework tutorials :shudders: )

I was aware of macports for quite some time (since I moved to macOS about 8 years ago) But most of the tools seem to be suggesting brew install, so I didn’t give macports a second thought.

8 years later, my MBP is old, and soon will not be supported by Apple and brew

While I upgraded to macOS monterey, this is going to be last version supported on this hardware.

After build problems with my PKM site, today I ran into another problem with this site.

What puzzled me was Deploy showed Build step failing but no useful error (other than exit code 2 - which does not help much)

But in the process of debugging, I learnt ¹ about the environment variable NETLIFY_BUILD_DEBUG (set to true when needed) It did not help 😞

I also upgraded hugo version (which is another environment variable) in the process.

In a multi-tenant rails application, usually we extract the tenant, before we process it further. Sometimes, the process requires us to process the data in background (via Sidekiq) and because the background job also needs to query the DB, it also needs to set tenant.

Turns out, we can pass “some” data from the controller to the sidekiq job via ActiveSupport::CurrentAttributes While, sidekiq may need to set the tenant again (for more DB queries) but one can save at least one query (to find out the account itself)