| 01:43 | <Kostya Petridi> | Hello, I apologize in advance if this is the wrong place to ask a question. I am wondering why the comment exists in the specification in the section about dialog description (see attached). Because a dialog is a single element that serves as a perfect base for popups, menus, tooltips, and any other piece of content that should be displayed over anything else on the page. Avoiding the use of dialog elements may leads to using workarounds (e.g., adding content directly at the top of the body), which might break specific scenarios (e.g., walkthrough elements in the tree). Am I missing something? Thanks. |
| 01:44 | <Domenic> | One small example of the many problems that can occur from abusing dialog for a non-dialog is that a screen reader will announce your tooltip (or whatever) as a dialog to non-sighted users. |
| 11:28 | <hsivonen> | annevk: https://github.com/whatwg/url/issues/110#issuecomment-261694624 indicates that WebKit switched to non-transitional IDNA processing in 2016. However, just now, I wrote fußball.de in the location field and pressed enter with the network inspector open in Firefox, Chrome, and Safari on macOS 14.3, and Safari is the only one that claims that the first request goes to fussball.de whereas Firefox and Chrome show the first request going to fußball.de, which redirects. Has WebKit deliberately switched back to transitional IDNA, is the network inspector omitting something, or do I misunderstand something? |
| 11:50 | <hsivonen> | TIL: Characters that are combining marks but that have canonical combining class 0, and UTS 46 is stricter about this than W3C charmod norm. |
| 13:37 | <annevk> | hsivonen: I suspect the inspector is omitting something. Maybe try something else that redirects? Bug reports welcome. |
| 14:30 | <Sam Sneddon [:gsnedders]> | it's also plausible that Safari is doing some processing of the location field before it reaches WebKit |
| 14:48 | <annevk> | It's possible, but if you try https://exampleß.com/ it gives a near identical error message to other browsers, so that seems less likely. |
| 15:08 | <hsivonen> | meßdienst.de works in Safari, so I guess the network pane in dev tools is what's weird. |
| 15:08 | <hsivonen> | Reading the old Bugzilla threads about transitional vs. non-transitional was interesting. I didn't recall that this stuff was so controversial. |
| 15:09 | <hsivonen> | Anyway, not that the three browser engines agree, should there be an effort to declare the transition over in UTS 46? |
| 15:10 | <annevk> | Overall I think it's dumb we did it, but then ICANN has been responsible for many TLD deaths as well so what's a couple of domain names at the hand of the IETF. |
| 15:11 | <annevk> | As part of my UTS46 feedback last year I did indeed ask them to remove the transitional mode. I'm not sure why they didn't. Might not hurt asking again, but perhaps we should have a more consolidated ask for changes as there are also test changes they need to make and such. |
| 20:52 | <Kostya Petridi> | One small example of the many problems that can occur from abusing dialog for a non-dialog is that a screen reader will announce your tooltip (or whatever) as a dialog to non-sighted users. |