Rust-Based Redox OS Had A Busy Year With Rewriting Its Kernel, Writing A File-System

[Gapil301]

Since minix has gone into the conversation, how is the project? Did Tanenbaum's son died?

[Nille]

they want to implement a FS written in Rust for their OS written in Rust.

They can Implement every FS in Rust if they want, but a new Filesystem is not necessary for this.

[Nille]

They said they tried ZFS. it is not that one should install Redox in late April 2017 for production use. Let them experiment what works and what not. There is no such thing as the single best Filesystem to simply choose.

ZFS is a bit extreme. But my problem with this decision is something else. You already tell it, its not a production use OS right now. So, if i want to access the Filesystem, i need something that is already in the Host OS/Developer Machine available of can be easy installed.

[jackpot51]

ZFS is a bit extreme. But my problem with this decision is something else. You already tell it, its not a production use OS right now. So, if i want to access the Filesystem, i need something that is already in the Host OS/Developer Machine available of can be easy installed.

There seems be be a misunderstanding here. RedoxFS (the current simplistic filesystem), and TFS (the future filesystem) both supply FUSE drivers, written in Rust with the same code used to serve the filesystem on Redox.

This means that Linux, OS X, BSD, and eventually the Windows Subsystem for Linux will all support RedoxFS and TFS.

[jackpot51]

ZFS is a bit extreme. But my problem with this decision is something else. You already tell it, its not a production use OS right now. So, if i want to access the Filesystem, i need something that is already in the Host OS/Developer Machine available of can be easy installed.

Both the current filesystem, RedoxFS, and the future filesystem, TFS, have support for mounting on Linux, OS X, and BSD using FUSE.

[mmstick]

The overall design seems right (the best off plan9 & minix) but targeting the x86 is already costing them in kernel purity with no real good reason other then the "Runs on Real Hardware" sticker... Well, whatever. Best of luck I guess.

Linux did perfectly fine targeting just one architecture in the beginning. There's really no architecture-specific code to be concerned about here besides the booting procedure though.

[Luke_Wolf]

Linux did perfectly fine targeting just one architecture in the beginning. There's really no architecture-specific code to be concerned about here besides the booting procedure though.

The other thing is the idea of not targeting x86 and instead a virtual machine is a bad joke. An OS targeting an obscure architecture on a virtual machine (which means we can't use Virtualbox or VMWare, and instead have to use QEMU) has a snowflake's chance in hell of taking off. Meanwhile at first glance, in all of a year these guys are further along than most of the Alternative OS projects, though I'll need to install it into virtualbox to check out.

Edit:
So having tested in virtualbox, further along than HelenOS, not anywhere close to ReactOS or Haiku

[starshipeleven]

Edit:
So having tested in virtualbox, further along than HelenOS, not anywhere close to ReactOS or Haiku

Should I take this as positive or negative? Is it better or worse than ReactOS (I find it hard to believe that it is worse than ReactOS)?

[lordnaikon]

I find it hard to believe that it is worse than ReactOS?

I find it hard to believe that you find it hard to believe that Redox, which is roughly one and a half year old, that is programmed in a programming language that is stable for the same amount of time is not as good as ReactOS, which is roughly 20 years old, that is programmed in a programming language that is roughly 40 years old.

[cj.wijtmans]

how about some TFS performance tests?