@zachleat wondering who's gonna give the next talk at @smashingconf 🤔
Subscribe
Public posts from @Schepp@mastodon.social
- Generator
- Mastodon v4.5.0-nightly.2025-10-01
- Public lists
- HWC DUS
Why is it that Chrome is the only browser that ignores inputmode="none" on inputs?
I want inputs in freshly opened <dialog> elements to not trigger the soft keyboard right away on (auto)focus, but instead when the user signals that they want to enter something, e.g. by a tap on it.
Safari works ✅
Firefox on Android works ✅
Samsung Internet works ✅
Chrome shows the soft keyboard ❌
☹️
Do I really have to inject a dummy element with tabindex="0" before the first input for it to catch focus?
Interestingly, inputmode="none" seems to work here, when dynamically applied: https://codepen.io/mustaqahmed/full/gOYxLaL
What's going on?
@KrijnHoetmer Happy Happy Birthday dear Krijn!!! 🎉🎊🥳🎁🥂 I hope you enjoy a wonderful day ☀️ with your girls - be it at the allotment 🌱 or at the beach 🏖️.
@chrisstoecker das kann doch kein Zufall sein. Ist das, um AI Crawler aus dem Tritt zu bringen?
If you really wanna fuck up tab order, I suggest you try to nest reading-flow into reading-flow into reading-flow...
Woah, totally doesn't work anymore! 😅
2. when I wanna programm a focus trap, like a modal creates one, it is suddenly not enough to do a querySelectorAll('iframe, input:not([type="hidden"]), select, textarea, button, a, [contenteditable], [tabindex]:not([tabindex="-1"])). I now have to select all elements and filter them for being scroll containers by issuing a costy getComputedStyle(). it would be better if there was a :scrollable pseudo class or similar.
@Chrome Q: starting with Chrome 132, you folks made scrollers focusable when they do not contain any focusable children. In practice that yields two problems for me:
1) to make certain areas not scroll through when a backdrop is visible, I create a pseudo element which cacthes scroll by makeing it scrollable and setting overscroll-behavior: contain. Problem now is that the parent element gets focus, even when it has focusable children itself (only the pseudo hasn't)
@iandevlin 👋🏻😊
@freddy hey and good morning! For some reason I came to think about the Sanitizer API and I wonder what the current status of it is. Do you know?
Had a wonderful day attending A-Tag in Vienna. The (accessibility) talks were fantastic and it was a great pleasure meeting old and making new friends 👏🏻😍
@MadeMyDay @sir_pepe WTF!
We all should use the term "pollywoppus" more often. Especially to characterize politicians.
En route to Vienna along the beautiful river Rhine. ✨
👏
> "Starting with Firefox 139, Nightly builds now support the hidden=until-found attribute, allowing content to be found via find-in-page when it's otherwise hidden by default." 🥳
@auphonic Und dann wollte ich noch fragen, ob Ihr den den Shownotes zugrundeliegenden Prompt vielleicht auch exponiert und editierbar macht? Denn für meinen Geschmack schreibt die KI zu viel im Passiv und auch generell nicht in unserem üblichen Tonfall. Das könnte man dann feinjustieren (+ Reset-Knopf, wenn man sich in eine Ecke gepromptet hat)
Liebe Grüße und vielen Dank Euch allen für Euren tollen Service!
@auphonic Wie wäre es, wenn man für jede*n Sprecher*in ein kurzes Audio-Schnipsel hinterlegen könnte, so dass das System anhand dessen eine Zuordnung vornimmt? Theoretisch könntet Ihr das Sample von jedem Sprecher auch nur einmal verarbeiten und das Ergebnis/den Fingerprint cachen.
@auphonic Guten Morgen Team Auphonic! ☀️ Ich hätte einen Feature-Request für die Transkription-Funktion. Und zwar ist mir letztens aufgefallen, dass ihr neuerdings die Speakernamen den transkribierten Sätzen voranstellt. Allerdings werden die in der Reihenfolge vergeben, wie sie in der Liste eingetragen sind und die Sprecher sprechen. Wir haben aber ein Preset, das immer übernommen wird und daher passt die Zuordnung fast nie.
I see some nice activity there on the Mozilla bugtracker for an implementation of `hidden=until-found` 😍
@Kilian It is also a lot more convenient than Chrome's local override feature and it would go hand-in-hand with Polypane's ability to disable CSP and CORS checks (for which I also needed separate extension, which are now also broken with Manifest V3)
@Kilian I do have a feature wish for Polypane:
Would it be possible to re-implement what the Chrome/FF extension "Resource Override" offered in the past, being that you could define URL patterns for website resources, which you could then remap to another URL or a local file e.g. https://polypane.app/**/*.js -> http://localhost:3000/**/*.js
I use that a lot to develop frontend-code, applying + checking it with the live-site, w/o having to run it via Docker locally.
👉 https://addons.mozilla.org/en-US/firefox/addon/resourceoverride/
Hey Ihr beiden! Gerade erreichte mich seitens @spittank folgende Frage zu meinem Post zur Verbesserung von "Find-in-Page" wie im vorherigen Toot verlinkt:
"Sehr spannender Ansatz!
Mich würde interessieren, wie gut das in der Praxis für Screenreader Nutzer funktioniert. Wenn ich damit teste fühle ich mich immer komplett überfordert."
Wie ist Eure Meinung zu "Find-in-Page" generell und zur Idee, Icon-Links oder Icon-Buttons per `hidden="until-found"` auffindbar zu machen?
Hey Rotterdam! Next week, March 4th, @fronteers teams up with Level Level to host a meetup. There will be talks about accessibility barriers in e-commerce & how embracing the web platform's native strengths is key to a solid design system. And there's still plenty of seats left! 👇