reflex VS webview

There's other people around here who would like to know your opinion about these GUI frameworks! I haven't written a GUI in Rust personally, but my favorite GUI framework is not at all OOP: https://reflex-frp.org/

Not to be confused with Reflex, allowing web apps in pure Haskell: https://reflex-frp.org/

FRP solutions sound very attractive. But reflex seems to be stuck on the outdated GHCJS, and I haven't been able to get it to build. The newer JS output in GHC doesn't yet have DOM support. And even if I used one of those, figuring out how to interact with a LaTeX renderer might be tricky.

I only have experience using Reflex, which I regard as the main contender for FRP UI libraries in the Haskell sphere. It's got a flashy website, but I think the documentation is a bit disorganized -- it took a long time for me to figure out how to get going with the library (you find some pieces of knowledge scattered here and there, if you look hard enough). My plan was to learn it well enough to onboard other people, but I don't think I could convince anyone who hasn't already decided that they're gonna make UIs in Haskell no matter the required effort.

Once someone has spent a bunch of time with Haskell and sees the value, they will find Nix if it makes sense. Maybe they'll want to play with https://reflex-frp.org, or they'll discover they want a better way to package 3rd-party dependencies, or they start using NixOS and want to nix all the things, etc. etc. Or, maybe they'll never find a use for it, and that's okay.

Reflex natively supports iOS, along with Android, desktop and web. I would recommend it for functional reactive programming in Haskell: https://reflex-frp.org

It's been awhile since I've tried to get into Ubuntu Touch/Linux mobile development in earnest. I'm currently working on an app using the reflex framework that I hope to eventually target Android, iOS, Desktop, and Linux Mobile.

If you're talking about the current Elm approach, I'm not sure. Otherwise, the paper I linked to notes some of the FRP libraries that existed at the time, some of which are still supported today (like reactive-banana), and otherwise I'd suggest looking at reflex, mentioned in the first post in this thread. I don't think it existed at the time the Elm paper came out.

You can create the webview using each platforms native GUI toolkit and setup JS communication yourself OR you can use a lightweight library that does it for [1] (search its README for language "bindings").

[1] https://github.com/webview/webview

Each platform has it's own webview control available as a shared library installed with the OS.

MacOS has WKWebKit based on WebKit.

Windows has WebView2 based on Edge/Chromium.

Linux has webkit2gtk based on WebKit.

Tools like Tauri use a simple cross-platform single-header abstraction called webview.h[1].

Electron no longer allows Node.js to be called from renderer processes, all communication with Node.js is done via IPC.

In this case, why do we still need Electron? Why does it have to be tied to V8/Node.js?

The fact that Chromium Embedded Framework exists and is third-party makes me think that Chromium wasn't designed for being embedded, and Electron is filling that gap.

This is elucidated here further here https://trac.webkit.org/wiki/WebKit2:

> it's difficult to reuse their work...if another WebKit-based application or another port wanted to do multiprocess based on Chromium WebKit, it would be necessary to reinvent or cut & paste a great deal of code.

It makes me think that perhaps WebKit was the better choice for embedding. The fact that Node used V8 made Chromium the choice, and that Node being called from the renderer was the original way of working. Maybe because WebKit didn't have a build for Windows was an issue too...

But now that we have Bun, perhaps it's time that WebKit becomes that browser target of choice for desktop apps on macOS.

Unless WebView2 for macOS arrives, which would have a more sane cross-platform story. WebView2 has a very large feature-set though which make take a while to implement for macOS.

[1]: https://github.com/webview/webview/blob/master/webview.h

Nui could base on this in theory. Nui uses https://github.com/webview/webview under the hood, which provides browser windows for linux, windows or mac. Nui adds some cmake to make the "in-browser" and "main-process" part appear seemless, as well adding a DSEL for the "in-browser" view part.

WebView 7k

You mean something like this?

Golang can compile to windows statically, and on Windows those bindings are using the MSWebView2 API (aka Microsoft Edge webview).

I know that you can also compile the webview.cc into a dll specifically, and link against that. But I'd never done with Visual C++ because I am cross-compiling from Linux to Windows.

The README of the webview/webview project refers to the WebView2 SDK on NuGet, however [1]

[1] https://github.com/webview/webview#windows-preparation

The author shrugs off web tech (maybe because of electron bloat?) but you can avoid the bloat by using each platforms native web browser control. There are even cross-platform libraries that make creating the native control and cross-communication simple. These applications would be architecturally similar to Win32 apps using and communicating with a XAML Island, but the advantage of web tech is it's an open standard and WPF/WinUI is not.

For example, you could use a native webview (Edge WebView2 for Windows and WebKit for MacOS/Linux), which uses much less RAM than Electron.

From a more established company with more guaranteed long-term support than the web frameworks that solve the above problems (like Tauri and Webview)