Notices by Nolan (nolan@toot.cafe), page 24

@SpankyWorks The only reason I write Bash is when I forget it's easier to just use Python. 😛

@codesections Basically in Pinafore I use LocalStorage only for small UI stuff like the current theme, the current instance, and UI preferences. IDB is used for offline storage of toots and profiles, which can be a bit larger than the LocalStorage limit and don't need to be accessed synchronously. Cache Storage is just for Service Worker-cached static assets.

@codesections

LocalStorage: Small, synchronous storage that isn't available in Web Workers or Service Workers. It's read into memory on page nav and limited to 5-10MB, so keep it for small stuff. Sync access means you can update it on page unload, though.

Cache Storage: good for content-addressable stuff (i.e. assets like JS, CSS, and HTML). Use in a Service Worker.

IDB: everything else. Large data, asynchronously accessed, agnostic to whether it's accessed on main thread or in a worker.

It seems Mastodon 2.6 changed how DMs work – they no longer show up in the Home timeline. Had to rewrite some of Pinafore's test suite to accommodate this. 😅 https://github.com/nolanlawson/pinafore/pull/630

I suppose it's incumbent upon apps to provide their own separate UI for DMs now. Overall though it seems like a good UI change. 👍

"Finding Our Digital Identities: A History of Social Media" by Jay Hoffmann https://thehistoryoftheweb.com/finding-our-digital-identities-a-history-of-social-media/

Interesting stuff. I wasn't aware of AsianAve or BlackPlanet. I never had a MySpace, but I remember how it seemed to spring up out of nowhere among my high-school friends. It's weird how so many of these things seem obvious in retrospect, even if at the time nobody knew how big "social media" would become.

HTTP-over-QUIC will become HTTP/3 https://daniel.haxx.se/blog/2018/11/11/http-3/

@chartier @sivy If there are security fixes I usually drop everything and do those. :)

@peter @sivy @chartier Yup, I do too. :)

The discussion is actually very interesting. Apparently the theoretical performance benefits of HTML Imports never really manifested in practice, either due to the one implementer (Chrome) not spending a lot of time optimizing, or lack of web developer usage, or both.

Given hacks like https://jakearchibald.com/2016/fun-hacks-faster-content/ actually being used in production in AMP, it actually does seem like there's a case for something like HTML Imports, but at the standardization level it just sort of fizzled out.

Chrome is going to remove HTML Imports in v73: https://groups.google.com/a/chromium.org/forum/#!msg/blink-dev/h-JwMiPUnuU/sl79aLoLBQAJ

@sivy @chartier Updated to 2.6.1 now 👍

@sivy @chartier Basically, yes. :) Also I am extremely cautious, as I've been burned by hopping too soon onto the latest Mastodon release in the past.