It's funny how Microsoft does a lot for backwards-compatibility of programs in Windows, yet it doesn't care at all about links to their documentation, release notes, or blogposts 🤦‍♂️

Seriously, I don't remember ever finding a link to a Microsoft article, that isn't broken after merely a year

Happy Day of the Programmer to all folks who consider themselves one! 🎉

This holiday was first proposed in 1996. The date of 13 September (12 September in leap years) was chosen since it’s the 256th day in the calendar year! 🤓

Version 2 of 🐘 toot, the cross-instance share page for Mastodon, is here 🎉

toot.karamoff.dev/

toot doesn't require JS to post any more and is easier to self-host (even without Vercel). The JS code that is left is now LibreJS-compatible. And, while hosted on GitHub, toot has mirrors on :gitlab: GitLab and :codeberg: Codeberg

For the first time in my life, I am developing a JS app without focusing on legacy. TypeScript codebase with compilation to ESM only. Minimal Rollup setup for bundling.

The best I've felt in my entire life :blobcathappy:

Last, but not least, implement stuff by yourself. This is basically like bundling, except you're responsible for the code that gets shipped. This adds lots of headache since one now needs to manage a way bigger codebase, but one also gets more control.

Show thread

Bundling dependencies + tree shaking can allow you to ship smaller bundles, but the dependencies will get outdated pretty quickly, unless you use dependabot or something to automatically push updates. Smaller size, yet less practical. And if one of your dependencies will have a vulnerability in it, your users won't find out

Show thread

Defining dependencies as dependencies has a benefit of not having to release a new version every time the dependency gets an update. The end user can run `npm up` and the newer versions will be downloaded automatically.

The con of this is the fact that everything gets downloaded. ReadMe, licence headers, in some cases even source code. And, of course, code that never gets executed in your CLI app. Kinda too much download size.

Show thread

Poll for the JS folk!

Suppose there is a JS-based CLI application. It's not that standalone, but rather used in other projects. Like a build system or a linter.

The CLI has some dependencies. Not a lot and not critical, but still. Stuff like colour and CLI manager, perhaps some concurrency library and other utility stuff

What would be your way of delivering it to the end user? I'll go over my personal pros and cons in the replies.

This is from the co-creator of Dogecoin (thread)

RT @ummjackson@twitter.com

After years of studying it, I believe that cryptocurrency is an inherently right-wing, hyper-capitalistic technology built primarily to amplify the wealth of its proponents through a combination of tax avoidance, diminished regulatory oversight and artificially enforced scarcity. twitter.com/ummjackson/status/

I really dislike it when application launchers sort the programs by categories or try putting them in folders. You know, like every Linux distro does by default. It's impossible for me to find anything because 90% of the time the categories don't make sense to me 🙈 Alphabetical order FTW

So, I have started using Brave Search since it, apparently, has the DDG bangs in it. Yet, the functionality only works when you _search_ with bangs, not when trying to navigate.

For example, "!yt linux" will open the YouTube Search Results, but simple "!yt" *will not open YouTube*, which is a shame :(

After 6 hours of running my bodged Python code on ca. 5k possible combinations, I think it's time to give up and just regenerate my 2FA revocation keys. Luckily, PGP and SSH keys were there for backup and are also present on my machine.

Show thread

There is a way to extract a hash from KeePass databases up to v3.1 and use Hashcat or John on them, which would save me tons of time. Apparently, similar stuff simply doesn't exist for KDBX v4 :(

Show thread

Lesson learnt the hard way: Always store your passwords. Yes, do it securely, use a password manager, hide a piece of paper behind multiple protection mechanisms, but DO KEEP A BACKUP OF YOUR PASSWORDS

Long story short: I decided to store every 2FA revocation code as well as every PGP and SSH key of mine in a strongly (64 Argon2d rounds) encrypted KDBX v4. As you may have guessed, I forgot the password. I do remember it slightly, but to no luck. Have been trying to brute-force it for 6 hours...

*a dirty mailbox

Can't imagine what a sorry mailbox would be like 😂

Show thread

I am becoming so ignorant of other people being ineffective. I get so aggressive when someone has a sorry mailbox, opens lots of windows, uses a mouse where a shortcut is faster and more logical. At moments like this I really just want to drop everything and start with "um, actually..."

I think I need a therapist 🤪

Alright, I've been advised (thx @kev and @0xedd1e)and to set up a PTR record for reverse DNS. Folks who use custom domain e-mails managed by others (Gmail, Mailbox.org, etc.): How do you do that? Do you contact the company? Do they do that automatically for you?

Show thread

I have set up SPF with DMARC as well as DKIM. I have never sent suspicious emails. I even add OpenPGP signing to my emails. SpamAssassin and GMail filters seem to not have any problems, when I check it with external tools. Yet still, my emails keep landing in spam folders of people unless they specifically add me to their contacts.

What could be the reason? Do email clients/servers not like email from custom domains? Or is it just because my email server is in Russia? 🙄

As a student of a university with access to a lot of scientific works, I would never think I'd have to use Sci-Hub until I graduate. But alas, Springer's institutional login is broken, and I really need that paper... :ablobcatwave:

Show older
Fosstodon

Fosstodon is an English speaking Mastodon instance that is open to anyone who is interested in technology; particularly free & open source software.