Got it. This is how federation works.

You need the local ID for the status you’re replying to. https://mastodon.social/@jkreeftmeijer/101236371751163533 is posted on https://mastodon.social, but you need the ID from https://mastodon.technology/web/statuses/101236371815734043 when posting a reply on https://mastodon.technology, for example.

I’ve used Vimari for a while, and I quite liked it. I don’t remember why it didn’t stick, though. @tombruijn, didn’t you use something like this?

Vim quicktip #3: Press gx to open a link under the cursor in your web browser.

You’re very welcome. This blew my mind when I learned about it. It seems to be one of those things you need to be told about. So, what’s a trick you can’t go without anymore?

I can’t seem to find how to post a response to a status on another instance via the Mastodon API.

Passing the ID as the in_reply_to_id parameter produces a 404 because the status can’t be found. That makes some sense, as that status isn’t on the instance I’m calling out to (right?). Switching the API base URL to the other instance gives me a 401, because I don’t have an API app set up there.

What am I missing here? Any pointers would be greatly appreciated.

Learning through reinventing the wheel

I love reinventing the wheel.

Some of my articles explain how to build your own GenServer in Elixir or how to compare images in plain Ruby, and I’ve built minimal clones of libraries like RSpec and Spring in the past to understand and teach how they work.

While you shouldn’t rely on a hand-rolled HTTP server or a naïve reimplementation of an ancient OTP construct in production, taking software apart and rebuilding it is the best way I know to understand what’s happening under the hood.

The results aren’t better than what already exists, or implemented in the fewest lines of code. That’s not the point. They’re built to be as expressive as possible to help explain concepts like HTTP, Rack, or inter-process message passing, and because they’re a fun exercise.

Randomly running into module :certifi is not available-errors in your Elixir project? It’s probably ALE’s mix linter recompiling your whole project, which is disabled by default in the current version.

https://github.com/benoitc/hackney/issues/528#issuecomment-444102939

The publication dates for articles on my website had broken <time> tags, so apparently search engines started using the update date as the publication date. Fixing the tags coincided with traffic from search engines dropping by half. 🤷‍♀️ #reverseblackhatseotips

Published in 2016, but still one of my personal favorites. It explains a bit of TCP and HTTP, some Rack, and it finishes with 30 lines of code that can serve Rack::Lobster. ❤️🦞

https://twitter.com/AppSignal/status/1072493054953709568:

@ghickman

$ ed
i
Ed is the stadnard text editor.
.
/stadnard
Ed is the stadnard text editor.
,s//standard/
,p
Ed is the standard text editor.
wq ed.txt
32
$

You’re right. That “all” wasn’t supposed to be there. :s/all/its/ 😅

Ha, necessary 18-byte evil, I guess. I tolerate borders for aside blocks (https://jeffkreeftmeijer.com/vim-16-color).

Also, aside from tables (which have a default border color), 1px outset is one of the smallest borders you can have (aside from ridge and inset, which save one byte).

I’m bringing border: 1px outset; back. 🎶

Same. As in; I keep falsely assuming neovim and Vim 8 are basically the same. I’ll start testing these tips in Vim 8. 🎩👌

I didn’t, but you’re absolutely right. It’s in $VIMRUNTIME/plugin/matchit.vim in neovim 0.3.1, which loads it by default (:echo exists("loaded_matchit") returns 1) while Vim 8 doesn’t. Like you said, it needs to be explicitly loaded with :packadd! matchit.

I don’t think you have to, as matchit.vim is included in Vim itself since 6.0. Your language plugin should set b:match_words for files of the correct type, so :echo b:match_words should return a list of match words if everything’s set up correctly.

Vim quicktip #2: A substitution with an empty search pattern replaces matches from the last search. After finding “foo” (/foo) in the file, you can replace all occurrences with :s//bar/.

TIL: A five minute lightning talk is too short to trace vim’s history back through vi and ex to ed while demoing both ,g/re/p and ,s/foo/bar/, so I renamed it to “ed: editing text like it’s 1969” and basically only showed how terse and user-hostile using ed is.

It would be great to use Twitter as a mastodon instance, allowing you to follow their users like you can with Mastodon users from other instances. However, Twitter will never allow that, seeing as they even forbid mentioning their users when reposting from other sources.

Unpopular opinion: in order for Twitter users to actually switch to Mastodon, the switch needs to be seamless. While reposting to Twitter might not be desirable in the end, it will get people to take the plunge.

because ed is the standard text editor!

That, and I’m giving a five minute presentation about why Vim is so weird tomorrow.

$ man ed

#currentstatus

I, for one, like my single-team Slack experience so far. Mentions in other teams arrive in my inbox after a while, so I keep track of one team, which is more than enough.

Accidentally added the wrong H&F team, however, so I’m in the same boat. Do share if you find that .plist!

ha, I literally can’t open, edit and save a file in Nano.

My whole setup is basically tmux in a terminal with Vim as my editor. tmux allows for windows and tabs to be able to keep projects and files separate.

Used Atom for a bit, too, by the way. Loved its plugins and actually-sane defaults.

Vim’s modular editing is the quickest way to edit and move around files I’ve used, and I prefer doing all development work in the terminal.

But yes, we’re mostly pretending. That message was Vimspeak for “you can use % to jump between brackets”. I might have overdone it with the backticks. 😬

So, what’s your favorite editor?