-
Notifications
You must be signed in to change notification settings - Fork 1.7k
Consider building static MUSL binaries #13086
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Comments
Maybe try to use the same build pipeline as rustc? |
You can already install rust-analyzer using
Not because GitHub removed a runner, because some of the users don't want or can't upgrade their distro. GitHub is not at fault here. |
@bjorn3 do you know if there are some preferred ways to name targets with statically- or dynamically-linked |
Not that I know of. |
As one of the louder people to complain in the other thread, I want to soften my stance a little bit: I quite like the solution where the It's really easy to install and when it is in the I think the What I am worried about is that this setup will also break regularly: usually our dev container is getting updates after a couple of Rust versions. This means in total the stable rust-analyzer could be 4 Months behind the vscode release. I guess sometimes the extension will already have evolved far beyond the stable channel Also the error message for the failure of the bundled version on an old system should definitely be improved, if this is an officially recommended way to solve the issue. |
@kamulos it's totally fine to run rust-analyzer from a nightly toolchain. But you can also disable automatic updates on the extension. |
From a purely practical standpoint, this just does not seem great to me:
I am just trying to find a pragmatic solution for my situation. Companies with slower updates will certainly always be there, and having a smooth, recommended way for us would be really great :) |
What about trying cargo zigbuild to build against an older glibc on the new image? |
musl don't support But rust-analyzer need |
Our
I knew that, so I don't know what I meant by opening this issue. Maybe I was thinking of a version that works except for the proc macros, but that would be terribly confusing for users ("I got the static binary because I love musl, but it doesn't work with serde"). So no, this seems like an awful idea. |
Our MUSL version is dynamically linked (for Alpine), but some users might just want a binary that works more or less everywhere. Perhaps we should build that, too (not sure under what name/target tuple), and maybe use it in the VSIX.
Some benchmarks on
analysis-stats self
, best of four:I didn't set
codegen-units
, but it looks like LTO might not be worth the trouble.The text was updated successfully, but these errors were encountered: