My first ever experience running rust software
>Windows>Rust>ROCmwhst the fuck is this unholy combination
>>108357247give swift a try, they are a security focus Lang but without being a bunch of tryhards
your fault most likely
>>108357247you were already running rust code long before if you browse the internet
>>108357247you see, a non-retard (depends on package "non-wintard") would have already obtained a backtrace (trivial), found out where a non-checked subtraction caused a wrap-around in a slice start expression (`slice_val[a-b..`, or `st= a-b; slice_val[st..]`), and then found out if it's a genuine bug or an error condition in your environment that could use better checking/error messaging (e.g. your disk was full).these happen to be the exact steps you would follow if this wasn't rust code, except you may not have gotten an immediate safe panic, and a very useful (to non-retards) message that you could screen-cap.
>>108357247If you want to use ROCm the best option is to install any Linux distro of your choice, it can be on a removable drive if you don't want to overwrite windows on your internal drive, then install docker and use the official images from AMD exactly as they instruct here.https://rocm.docs.amd.com/projects/install-on-linux/en/latest/how-to/docker.html
>>108357345>>108358766rocm is irrelevant to the error here. the error condition hit which i explained in >>108358746 seems to be in uv, which (i'm assuming) happened during installing python packages via it, where rocm happened to be one of them. the only part where rocm could be relevant is the fact that the package is sort of atypically huge. that's it.
>>108358804You can't expect uv, python, rust, ROCm, etc. to work on windows, windows is for microsoft office, photoshop, video games, and copilot.
>>108358836You can't expect rust developers to grow up, put on their big boy pants, and develop for the platform where most of the users are.
>>108358918The reality is most AI users are on Linux, which is why nvidia's Windows driver has been breaking constantly because they moved all the best devs to work on the Linux datacenter driver instead.It's easier to just boot into Linux for AI then go back to Windows when you're done if you don't have the resources for a dedicated AI server.
>>108357345kek this
>>108358836i don't know about rocm, but i expect uv is very much expected to work on the wintard platform.nothing super mysterious is happening here. i already outlined how the wintard op can get to the bottom of this in literally a couple of minutes.
a language should be designed such that it is impossible for unhandled runtime errors to happenevery possible error must be handled in source or it should throw a compile error
>>108359850sounds like a great way to make a language that can't do anything
>>108358104Glad to see you didn't fall for the marketing of a company renowned for its marketing, apple boy.
>>108358918I don't give a fuck about Windows users, they have WSL anyway, they can run shit there.
>>108359914it's not hard to handle runtime errorsmaybe in Rust it's hard but that's a Rust problem
>>108357247>R-ACK-m
>>108359927its open source so whats your point?
>>108360034By blowing up yeah, which is what Rust is doing there"Handling" every possible error in the sense that you can recover from it is not easy or even possible. You would need 50000 lines to create a file and the result would possibly nuke your system
>>108363004Yeah it's just a difference in comprehension I guess. The user shouldn't ever see a panic, simple as.
>>108357247virgin rust vs chad golang
>>108363753So the word panic is your problem? In C the same problem would give you>segmentation fault (core dumped)which I guess is okay thenwhat's funny is I do see a real problem with the screenshot which is that it only gives you the offending line and not a traceback, but my guess is you would complain even more about seeing a ugly stack trace. I know your type. Enjoy your>oooopsie we made a fucky wucky!! :(
