Nyxt 2.2.0 | Hacker News Reader

Nyxt 2.2.0

nyxt.atlas.engineer

152 pointspedrodelfino35 comments

I use Nyxt as my daily browser for the most part.

After experimenting with lots of alternative browsers (uzbl, qutebrowser, etc.), I find Nyxt to excell in the following areas:

- Configuration system is highly organized as well as flexible;

- Feels like a first-class browser in almost all situations;

- Fast startup and responsive UI.

That said, I've encountered some issues, mostly in connection with work:

- CJK input is kind of janky;

- Infrequent crashes on some JS-heavy sites;

- Cannot log in to Azure Data Factory for some reason.

All in all, I would highly recommend giving Nyxt a serious spin. The devs are crazy active and helpful. I am pretty confident that the above issues will eventually get ironed out.

I wonder why does it crash. It uses WebKit as renderer. Can you give an example of a page on which it does crash?

Unfortunately, I have been just ignoring these lately, with the intention of troubleshooting later. Now seems like a good time to at least start taking notes of problematic pages. Cheers.

No binary distribution available for about 98% of the world's desktop computers and they only linux distro they make packages for that has anything approaching significant desktop share is Arch.

It's like they're purposefully being snobs making it only available for the most obscure operating system users, or something.

We have a deb installer. We are trying to make it as available as possible, but we don't have the computers necessary to test and distribute on those platforms. Sorry about that!

That's great news, but if it's not listed on the download page, people aren't going to know that.

> we don't have the computers necessary to test and distribute on those platforms

Even if hardware wasn't cheap and virtualization hasn't been readily accessible for over a decade, if you're constrained on hardware why are you focusing limited resources on packages for incredibly obscure linux distros? That are more difficult to develop on because a)they're not as high quality or stable and b)there are less support resources and documentation for them?

Re: OS X and Windows, cross-platform support is just going to be all the more painful later.

It's really not that hard to get a working windows-installation for low money, if you actually care. Nowadays, you don't even need a license it seems, just install the official images into a VM.

If you want to invest some money, there are also cheap refurbished PCs, cloud-desktops or some less legal sources for a license from eBay or Craigslist.

If you believe it's now worth the time and effort to make such binary distributions, I'm sure the devs would highly appreciate your help and compute power.

This is extremely common for new open source software. You can find plenty of interesting projects only packaged for Arch and Nix, because packaging for those distros is extremely easy. It's just a matter of time and effort, and this is a volunteer project.

I really want more firepower and ergonomics out of my browser. Yet I have not found a better solution than Firefox + Tridactyl + a bunch of other extensions. Two strong contenders are Nyxt and Vieb right now but neither checks all the boxes below.

I want

- complete Vi like keybindings - almost everything should be doable via the keyboard

- window management - why can I just not split a browser window into two?

- fast and powerful engine - I don't want to use two browsers

- Webextensions support

It's a bit of a mess, but have you tried using emacs-webkit[0]? You can have vi-style keybindings via evil-collection, the window management is pretty good, the browser engine is pretty standard. I've used it with spacemacs, and it's overall a decent experience.

It's not quite Gecko though, and I don't think WebExtensions work yet. Also the single-threadedness of emacs itself can be a bother (of course the browser component is separate).

link to emacs-webkit: https://github.com/akirakyle/emacs-webkit

This always fails with Webextensions support. The simpler solutions just take an existing webengine-component from gtk or Qt and put a browser around it, but this fails with more complex features which are naturally not part of the webengine.

I'm curious why nobody takes directly chromium or firefox and just integrates trydactyl to get first-class-support. The major-problem with alternative keybindings is the context they run with. Extensions run inside the page-scope, and thus are only available after the page loaded. Integrating them at app-scope (which mozilla promised to deliver around 4 years ago already), would basically solve most problems instantly.

Yes, it is incredibly hard to build an extension system.

Why does nobody build on top of Firefox and Chromium? Probably because tracking upstream changes will be incredibly difficult. I imagine as soon as FF releases a month worth of patches, trying to integrate them and sift through them would take thousands of hours.

Webextensions support is /very/ close, we even have a pull request for basic support: https://github.com/atlas-engineer/nyxt/pull/1762

now that we've figured out the framework/setup we just have to implement the APIs.

Documentation improved a lot in the recent versions. It’s still a WIP, but I had a ton of fun enjoying emacs-like philosophy natively supported in a browser.

Neat product, but the lack of extensions is my main problem. I've been using Vimium with Chrome for keyboard navigation and it works well without losing all of the benefits of Chrome.

https://chrome.google.com/webstore/detail/vimium/dbepggeogba...

Not sure why you're only mentioning a vim extension when that's exactly what Nyxt is, but better in every way.

Are there other extensions you want that are missing in Nyxt?

I want all of the features of Chrome like extensions, profiles, syncing, passwords, etc. That's why I mentioned Vimium; because it brings keyboard nav to Chrome, instead of leaving Chrome for Nyxt.

It would be great to have uBlock Origin, uMatrix, firefox's containers and other privacy features.

If have nearly 50 extensions in Firefox. Nyxt will never implement them all. That's why APIs exist, to allow people doing it for themselves.

I used nyxt but the lack of extensions really stopped my use of it dead in its tracks

I need a password manager for instance, and its nice to have it integrated.

Having bookmarks sync'd too is a requirement.

Firefox containers would also be nice too since I want more control over privacy.

Indeed, that is why we've been working really hard at it!

https://github.com/atlas-engineer/nyxt/pull/1762

we are incredibly close.

Looks like this update solves all the issues I've been having using it as a daily driver. The input stutter was incredibly bad sometimes when you have muscle memory and had to intentionally slow down. Gonna do a guix pull!

Would be nice if the project could support Gemini and Gopher as well

There are Common Lisp client libraries for Gemini and Gopher, so I imagine it would not be that difficult to add support.

Wow, looks great. Might replace my Firefox with aging pentadactyl extension.

Small typo on front page: "that allow you to quickly analayze -> analyze"

You can use tridactyl, a well-maintained and active project for vim bindings in firefox: https://github.com/tridactyl/tridactyl/

I’m pretty psyched to see the progress being made on this project, keep up the good work!

not available for most popular OS

Pity there is no brew package.

I second this. I like Nyxt but the macOS install experience can be painful.

Indeed, I wish there was an easier solution for macOS users, in the meantime you can try https://github.com/deddu/nyxt-docker

kudos on the new features

Nuxt, Next, Nyxt. I can't keep up.

Reply on news.ycombinator.com