@musicmatze no there is no way. I have to be on the bleeding edge... Can't resist the temptation... 😂

@musicmatze I might be lonely in that opinion, but everyone should just always use the latest version and upgrade whenever possible.

I personally would even bump the MSRV in a patch release for any of my crates simply because of that.

I understand not everyone can always use the latest version if on some very special hardware but in general upgrading is a no brainer, especially with rustup.

@ayushsharma22 nice thing about is that you can easily switch the cloud provider with and try out all the provisioning steps locally.

Sure, it's not 100% the same as the base image is different so you still can expect one or the other thing to fail when running against the cloud provider.
But still it's such a helpful tool to develop the setup all locally first.

@ayushsharma22 thanks, I'm already using Packer.

Using a mix of Terraform and packer for my deployment and used packer with shell scripts so far.

Now had a little chat with a DevOps guy who uses Ansible a lot and after that I finally started trying it out now.

With shell scripts there are just so many things that can go wrong, often failing my provisioning.

Just discovered for me. Of course did know about it but never properly checked it out.

Such a nice tool and just great to provision a base image for VPS instances. Especially not having to fiddle with all the edge cases when provisioning with shell scripts.

Still much to learn, I feel I just scratched the surface of it... Well, my setup is rather simple anyways 😅

@janriemer this is a neat tool I didn't know about (and my cargo installation list is already huge 😅).

Sure, it's not a safe tool for guaranteeing API compatibility as already pointed out by others.

Still, it gives a nice overview of the API changes, not just as maintainer but also as user of a crate. Especially if maintainers don't keep a changelog.

@musicmatze yeah that some messy AI assistant that gives you all the wrong input all the time... But I don't see why that would directly turn you to jump off the GitHub platform with all your repos 🤔

@musicmatze what did they do that you want to move all your repos out of GitHub?

@anahata cagro build

Happens to me quite often when typing quickly 😅

@tobtobxx especially the recently added msrv feature, dependency changes in the version history and distribution of version usage. All very useful features/graphs that crates.io doesn't have or presents in a much less readable way

The video of our first steps into implementing "A Practical Wait-Free Simulation for Lock-Free Data Structures" in #rustlang is up! I miss these longer streams — they're a lot of fun. Still some way to go, but we made good progress! youtu.be/Bw8-vvtA-E8

Just lovely ❤️: "However, note that Rust is a project run by volunteers. We prioritize the personal well-being of everyone working on Rust over any deadlines and expectations we might have set. This could mean delaying the edition a version if necessary, or dropping a feature that turns out to be too difficult or stressful to finish in time."

Go have a look at the blog post, if you want to know what's coming for edition 2021.
Very impressive, how they plan to fix some problems.

@musicmatze @rugk @skyr I feel the only good reason to do conventional commits is for either huge projects with quick release cycles to keep it manageable (although even then there are many that can do without).

And people that just don't care about their changelog much and are too lazy to write proper ones, instead generating them.

Sorry if that comes as a bit spicy. It sure can have value but mostly is not really needed. Same opinion as @musicmatze there.

@kirschwipfel @txt_file ... Dinge die direkt in den Sinn kommen sind Generics, Feature Flags und Compiler Flags.

Generics können nicht als shared lib kompiliert werden da sie zur compile time joliet werden wie Templates und anschließend optimiert. Deshalb sind sie sich nicht in FFI verfügbar.

Feature und Compiler Flags können die erzeugte binary und enthaltene Funktionalität stark beeinflussen. So kann keine lib kompiliert werden die "für alle" funktioniert.

@kirschwipfel @txt_file Das statement klingt aus zweierlei sich fragwürdig.

CI/CD kann auch in Rust viel caching verwenden, sodass sehr viel inkrementell kompiliert werden kann, Test Ausführungen werden momentan leider noch nicht gecached. Klingt daher eher nach fehlerhaftem Setup wenn jedes Mal alles von scratch kompiliert wird.

Das andere Problem ist, dass Software Bibliotheken nicht so ohne weiteres in Rust als shared lib kompiliert werden können.

@musicmatze this looks really interesting and is perfect for first time setups of a cli app.

Was just about to say, this screams for a proc macro but seems you already have that on your to-do list 👍.

I'll play around with this nice crate later.

You prefer PRs/changes by email iirc?

@kornel you're rocking it! lib.rs is really great and replaced crates.io for me quite a while already.

Every new feature that came in the past was really great, like the dependency graphs for example.

Would be great if cargo added set msrv setting by now so you don't have to use so much computing resources.

Keep it up, your work is great 🚀

@wholesomedonut it's interesting that they make such a big deal of it for Rust but with Golang for example, nobody seems to care.

Same as with Rust they add absolute paths of the files to print them during crashes. Not sure whether there's an option to remove that through.

But the point is, Go has quite a bigger user count and there nobody seems to care 🤔. And probably many other languages do the same.

Nonetheless it would be nice to have an option for this in the future.

Show older
Mastodon for Tech Folks

This Mastodon instance is for people interested in technology. Discussions aren't limited to technology, because tech folks shouldn't be limited to technology either!