Hacker News Reader

Comment by neoromantique | Hacker News Reader

Why? You can treat Dockerfile as documentation for the most part

Unnecessary complexity makes debugging and understanding the system much harder.

This is particularly common with CLI tools written in some languages. I was looking at Antora the other day (not intending to single this project out, it's just the one that came to mind). I found two ways to run it:

1. By installing Node: https://docs.antora.org/antora/latest/install-and-run-quicks...

2. By running it in a Docker container: https://docs.antora.org/antora/latest/antora-container/.

The amount of complexity here is shocking. This is a tool that could just as well be a single binary, with the only dynamic linkage being to libc and maybe OpenSSL.

This also means that if something goes wrong, black-box debugging tools like system call tracers are much harder to use. I rely on system call tracers all the time, and it really sucks when they stop working.

It's just "complexity" if you aren't super comfortable with docker.. it's super easy to do everything you describe with docker. Like for me it's much easier to debug in a self contained system, because even a binary can have issues with dynamic linking, etc. So for me the complexity is reversed. I don't want to pollute my actual machine with stuff when a docker container is just as easy to use. I don't want my distro's OpenSSL to be slightly incompatible with something that the package is using. A dockerfile removes all of that.

Well, distributing CLI tools as Docker images came about in part due to environments like Node, which made it harder to ship a single statically linked binary.

Imagine, like, your source control tool being shipped as a Docker container.

I agree yes, CLI tools should not come in docker packages. I'd also blame python for that, it's harder to package.

Don't get me wrong, I absolutely would love for everything to just be statically linked and packaged in a single binary (an approach that works great on windows usually). And you are right that "over using" docker is kind of a trend, but I think it's due to a problem (packaging apps in Linux) rather than being a problem by itself.

Because despite most devs opinion on the matter, we don't live in a Linux only world.

you have access to docker on both windows and macos

Under virtualization (or emulation if amd64 on arm64). May as well spin up that VM.

I'm not sure what your point is. virtualized or not you can run docker on any mainstream operating system using any mainstream hardware and get near native performance.

Outside of development, running containers on macOS/Windows doesn't make sense. And macOS is using emulation via Rosetta, not virtualization on M-series.

Only if there is no arm variant of the image you want to run.

Your Linux VM instance is Linux, and I don't think it's an unreasonable request to run a VM on your desktop machine, using the virtualization software provided by the OS.

Reply on news.ycombinator.com