http://example.qt.io/qt-webassembly/widgets/richtext/textedi...
http://example.qt.io/qt-webassembly/widgets/richtext/textedi...
It's like the web version of godawful cross platform Java apps from the 90's where they stick the Mac version's menubar inside the window because it looks right on their machine.
I wonder if screenreaders work for reading and navigating this.
EDIT: This comment comes across as totally negative, so I'd like to amend that I think there are good use-cases for it. Games or more visual things, sure, I don't care if the text isn't a pixel perfect match to everything else on my computer.
But for things where reading or writing text is the primary purpose, having the same font face at the same size render differently from the rest of my computer is not something I want to see becoming common.
Dude, that battle was lost when Electron gained traction. I hate it with all my heart, but it's a sad reality.
Good font rendering makes SO much difference in how stuff looks. No joke: font rendering is the number one reason I prefer the look of macOS over Windows (especially now since WSL exists). I realize I'm a weirdo font nerd and probably alone in that stance, but working with computers mean staring at text on a screen for hours and hours every day. It fucking better be rendered right.
There are other issues like contextual menus, in VS Code they're handled well and give you all the actions you'd expect. But in Slack it gives you the bare minimum OS included items and has their own "..." button on each message for things that ought to be in the contextual menu.
It can also do multiple windows (again, see VS Code which is a good example), but Slack stuffs everything into one window (preferences overlay is a good example of what should be separate) because it's a much simpler wrapping of a webpage instead of being designed as a native app would be.
Unless Webassembly provides a common type rendering system or links to the Browser's host renderer, I'm not sure this can be fixed.
It's just pixels. Barring patents a wasm program can text any goddamn way it wants to.
Sure there is - put the text in the DOM.
WebAssembly is just a runtime, it has nothing to do with whether you display your stuff in Canvas or via the DOM.
If you were writing an app in Javascript, you wouldn't bundle a text engine into it and then bypass the DOM for all of your rendering. Why is WebAssembly different?
Sounds like a good proposal for the WHATWG.
Overall, I think that's a decent trade-off. Most of the UI is quick, easy, and (potentially) extensible using HTML while the performance-sensitive parts move to what's basically a raw buffer.
https://code.visualstudio.com/blogs/2017/10/03/terminal-rend...
For reference: http://jakiestfu.github.io/Medium.js/docs/#invoke