In homelab, I push myself to use proxy (header) authentication. I know I'm burdening many responsibilities in a reverse proxy (tls, ip blocking, authentication) but it seems I can better handle those complexity as compared to oauth setup.

> A UDP joke I would tell you... but you might get it not

Except the text is stored as image, I can't imagine localization would be that large.

I'm sure no API and only built-in control is more favorable. Digressing, built-in mixer is nice to have too.

experiencing how text renders differently, slowly, with my potato battleship

One other angle yet mentioned: JS is browser native. No matter how slow it is, browser is now the LCD. Similar server-client codebase, while ugly, is another plus.

I think the hack is to store html height/width locally and restore it as early as possible so the content will then load under the scrolled view

Oh hey, a fellow noticing person!

yes, and the bar is not at all at the same level.

Now ui libraries does its own reset. The no-nonsense blogger just use a few styling, and yeah, the default now is good enough.

Back in the day I requested chrome feature "copy text" in addition to "copy link" on <a> context menu. Now I tried it it's no longer there.

It's not? Just checked a Chrome instance I had handy, it has all three options in the context menu - "Copy", "Copy link address" and "Copy link to highlight". First one copies text in between <a> ... </a>, second one copies the href attribute, and third one copies the link to page you're on with that weird URL framgment-based arbitrary text anchor/highlight scheme.

All three work on Google search results for me.

Maybe weird behavior on my end? Or perhaps you need to select part of <a>'s content to trigger it?