jdev - 2026-01-07

  1. kapad

    display `ul` and `li` as inline then `p` really make sense. i personally dont even use `head`, the 1st non-head element and the body just started...

  2. kapad

    `doctype` and `html` are enough for a valid html5 page.

  3. theTedd

    Browsers have to be pretty lenient in what they accept, otherwise a lot of web pages wouldn't work very well; for this reason, "seems to work in my browser" is not a good measure of "a valid html5 page"

  4. singpolyma

    theTedd: it's not just leniency or "works in my browser" anymore though. Quirks mode is dead. HTML5 defines exactly what to do in every case so even if it's "not valid" it has a single, defined outcome and will be the same in every compliant browser

  5. theTedd

    What percentage of the web is html5? What to do with the remainder that isn't?

  6. singpolyma

    You mean if it explicitly declares html4 or XHTML1 and then isn't compliant? Yeah then I guess who knows in that case.

  7. theTedd

    There is a lot of non-compliant (X)HTML and browsers still need to do something with it; if they just fail then the browser is broken from the user's perspective, so they use a different browser that doesn't fail. Back to my original point: "seems to work in my browser" is not a good test of validity.

  8. singpolyma

    Sure. I guess I didn't consider intentionally Non compliant legacy content. I can't imagine it's a big % vs unmarked tag soup but I'm sure there is some out there.

  9. theTedd

    It's more a case of not knowing the spec than intentionally not following it, plus "it seems to work in my browser, so that must mean it's correct"

  10. singpolyma

    At least there wont be anything new created in that category so the % will drop over time pretty fast I expect

  11. singpolyma

    I don't know when I last saw a legacy doctype on anything

  12. и

    иии

  13. jjj333_p (any pronouns)

    i have a jank bot that produces html pages just as output from a python markdown parser i dont even add a head or any of the proper html page things. renders fine though id like to add an actual page to wrap the content in eventually for aesthetic reasons

  14. jjj333_p (any pronouns)

    ive yet to find a browser that chokes on it. firefox complains in dev console but it doesnt choke

  15. singpolyma

    head is not required in the spec anyway

  16. singpolyma

    nor html tag

  17. jjj333_p (any pronouns)

    such html: https://downloadable.pain.agency/file_share/019b9a6c-a83e-7a53-9aba-db56d1a55a23/o.html

  18. jjj333_p (any pronouns)

    ignore the content its whatever

  19. jjj333_p (any pronouns)

    it reeks of markdown to html lol

  20. singpolyma

    that's almost valid. you need a <title> officially (doesn't need to go in <head>) if you add that and an html5 doctype it'll validate

    ❤ 1
  21. fpyrih

    > I'm trying to figure a way to keep E2EE through a bridge if the other end also implements Signal Protocol for E2EE (for ex., Matrix) do any of you happen to know if Signal Protocol generates homomorphic cyphertext? I'm trying to figure if I can take an OMEMO 0.3 cyphertext and "envelope it" in JSON without decrypting it so that if msg_plaintext is "Hello world!", `json_cyphertext_begin + msg_cyphertext + json_cyphertext_end` turns into something like ``` { "type": "m.room.message", "content": "Hello world!", } ``` when decrypted