A Web Framework for Zig

159 points by nivethan 248 days ago | 23 comments

lgrebe 248 days ago |

Jetzig reads funny in german something like "now-ish"

Jetzt = now [1]

German adjective suffix: -ig [2] The German suffix -ig attaches to nouns, verbs and even adverbs. Given this flexibility, it ranks among the most common adjective endings in German. You can use -ig words to express that something is a certain way or happens a certain way. traurig: sad, sadly wässrig: watery knackig: crunchy, crispy abhängig: dependent, addicted geizig: stingy

[1] https://de.pons.com/übersetzung-2/deutsch-englisch/jetzt [2] https://www.lingoda.com/blog/en/german-adjective-suffixes/

unwind 248 days ago | |

The "-ig" suffix corresponds roughly to English's "-ish", I would say. As a native speaker of neither German nor English, though.

rendx 246 days ago | |

It doesn't merely sound like it, it is a German word, translating to current/present.

https://www.dwds.de/wb/jetzig

https://de.wiktionary.org/wiki/jetzig

https://www.duden.de/rechtschreibung/jetzig

Onavo 248 days ago | |

It's a perfectly fine name. There's a billion dollar web framework company named Zeit which famously built https://now.sh

moritzruth 248 days ago | | |

They rebranded to Vercel, most likely because Zeit turned out to not be a perfectly fine name.

endigma 248 days ago |

This site makes a big point of using the term "RESTful" repeatedly, but it seems to be JSON-based by default?

I don't know why modern web frameworks insist on continuing to misuse or misapply the term despite a fairly large amount of messaging recently about how exactly this term is misapplied, and the resurgence of frameworks and tools that do correctly apply it, e.g. HTMX, Datastar, Alpine AJAX.

Otherwise, this looks cool. I'd encourage you to un-roll-your-own docs and use something like Starlight or Docusaurus so you can have usable search and versioned docs.

_heimdall 248 days ago | |

That's a fight we lost two decades ago now unfortunately. Nearly any modern-ish API is a JSON-based RPC. There's nothing wrong with that, JSON RPC is a plenty fine solution for many common use cases, it just isn't REST.

moritzruth 248 days ago | | |

Note that there is also a standard for JSON-based RPC systems, called JSON-RPC [0]. Not every JSON-based self-titled "RESTful" API uses JSON-RPC.

[0]: https://www.jsonrpc.org/

elcritch 248 days ago | |

Being RESTful and the data encoding used are largely orthogonal aspects of an API.

Wikipedia RESTful article says:

> The formal REST constraints are as follows:[10]

> Client/Server – Clients are separated from servers by a well-defined interface

> Stateless – A specific client does not consume server storage when the client is "at rest"

> Cache – Responses indicate their own cacheability

> Uniform interface

> Layered system – A client cannot ordinarily tell whether it is connected directly to the end server, or to an intermediary along the way

dustbunny 247 days ago | | |

The key line is actually this:

>An application that adheres to the REST architectural constraints may be informally described as RESTful, although this term is more commonly associated with the design of HTTP-based APIs and what are widely considered best practices regarding the "verbs" (HTTP methods) a resource responds to, while having little to do with REST as originally formulated—and is often even at odds with the concept.

jadbox 247 days ago | |

What do you think of HTMX vs Datastar for a new production app? Any key reasons to use one over the other?

dustbunny 247 days ago | |

Ha I came here to complain about the same thing.

Kinda makes me disappointed in the authors especially when they link to the REST wikipedia page as if they know what it means.

BozeWolf 248 days ago |

Nice, looks like a decent framework. I used to do a lot of python for backend web apps, but recently jumped on the hypetrain and used go for developing a web app (devops) tool. Single binary, easy deployment etc etc.

From that experience, I think this competes with go based web apps mostly. And if so, it makes a good chance at becomming succesful. Zig seams to have a better type system. Additionally the quality of documentation for this project is pretty good. That is something the go ecosystem seems to be lacking in general. The rest of the go perks are there as well. Single binary etc.

Now the ecosystem needs to catch up. Nice!

heldrida 248 days ago |

Here's a video introduction to Jetzig by one of the contributors:

Screencast #1 - Introduction https://www.youtube.com/watch?v=qeaO_MwfUII

sroerick 248 days ago |

I really like Jetzig. It just seems to make every framework decision that I think feels right. Major props to the team

timeon 248 days ago |

Sad to see that cookie banner. One would not expect opensource project to be associated with spyware.

lukaslalinsky 247 days ago |

As much as I like Zig, I can't imagine working on something as string/list-heavy as web without a garbage collector. Not to mention having a web server without coroutines. Web services these days are just glue code between various pieces of infrastructure, you are constantly waiting on something, it just doesn't make sense to run these in full blown threads.

winrid 248 days ago |

I don't know zig - is this query type safe in that if I change a member name in the struct it'll fail to compile? https://github.com/jetzig-framework/jetquery

usually I do this with codegen, pretty cool if zig's type system is this powerful

Also holy crap: for (cat.homes

This is like django!

messe 248 days ago | |

Yes, all reflection in Zig is static reflection, so that should generate a compile error (although how obvious it will be is a different question).

hajimuz 248 days ago |

it’s using http.zig, so only http1.1?