Notices by Lars Wirzenius (liw@mastodon.acc.sunet.se), page 2

Bye-bye X220, you served me well. Maybe I can diagnose and fix whatever hardware problem you have, some day.

Helsinki from somewhat high up

#helsinki #skywheel #midsummer

I don't like heights, so this was a mistake. But not too bad a mistake.

#helsinki #skywheel #midsummer

The two fundamental problems in computer science are cache invalidation; naming things; and avoiding off-by-one errors.

You have some code that works, but you need to add a small feature. You change the tests to verify the new feature works, before you add the feature, because you like Test-Drive Development.

At this point the tests will fail. Don't panic. It would be embarrassing to panic because the tests for the thing you haven't written yet fail.

(You should panic if they don't fail: in that case, your tests are buggy.)

I self-host my git repositories using #gitano (https://www.gitano.org.uk/). It does not have a web UI (I use cgit), and doesn't do pull requests, tickets, etc, but it does have a very flexible access control system, which I like.

I have an account on GitHub, originally to file a bug on Paramiko (still open). I now also have one project there, to experiment with the pull-request style of development (don't like it on GitHub). I was never going to use GitHub much, since it's a thoroughly proprietary service.

I don't like GitLab, either. They are not sufficiently free software for me to be happy (they're "open core" instead), but I understand why many others think they're free enough.

That toot was riggered by me writing "127.0.0:1", note the colon that should be a period.

Took me almost a kilosecond to debug. Too much stress -> bad sleep -> tired -> mistake-prone. Perfect time to write code, eh?

I keep failing to follow rule 1: "Always copy and paste a URL."

It applies to anything that's easy to mistype, really: URLs, email addresses, bug numbers, IPv4 addresses, IPv6 addresses, many domain names, passwords, some people's names, long command lines, ...

https://liw.fi/rules/

If your solution to the surveillance economy we live in, is to tell people to "just don't share" personal information, I feel you don't understand the world enough to talk about this issue.

@joeyh @n8 Interesting.

If you're aiming to produce a program or service that is highly peformant, start measuring early on.