Typing conformance results

No changes detected ✅

mypy_primer results

No ecosystem changes detected ✅

ruff-ecosystem results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

✅ ecosystem check detected no linter changes.

Formatter (stable)

✅ ecosystem check detected no format changes.

Formatter (preview)

✅ ecosystem check detected no format changes.

I'm going to put this back in draft for now. I'll be curious to get @MichaReiser's thoughts when he gets back.

FWIW #17584 also fixes this issue by resolving the underlying problem (I've just run a CI on ubuntu-slim with and without the changes to check)

Can you say more about why we have to fork is-wsl? Is it because it depends on is-docker? I'm not sure if it's worth it but we could consider creating an upstream PR to is-docker, they certainly have more background knowledge than I to review the change.

Yes, is-wsl depends on is-docker. Both are one-file dependencies from the same maintainer. Both were last updated 3 years ago, which is why I didn't try opening a PR upstream, but I certainly could try. Their READMEs both say they're Rust ports of https://github.com/sindresorhus/is-wsl and https://github.com/sindresorhus/is-docker from JS.

If we decide to fork the crates, it's important that we attribute the source of the code by including a header like this / Adapted from <crate> by <author>, MIT License

Ah okay, makes sense.

I'm also okay closing this if we'd rather not bother. I just got curious and found a fix but wasn't confident it was a good one.

I'm also okay closing this if we'd rather not bother. I just got curious and found a fix but wasn't confident it was a good one.

The fix itself seems worthwhile to me.

Sounds good, I'll try a PR upstream.

Yeah, is-docker seems more abandoned than is-wsl. Could you use cargo's dependency override feature to only fork is-docker (I don't think it's worth spending much time on but could be nice)?

I think I might have to create a crate for that? I'm not sure but I'll give it a try!

I think I might have to create a crate for that? I'm not sure but I'll give it a try!

Oh yeah, that might be more work than it's worth

Memory usage report

Memory usage unchanged ✅

I prepared a patch for is-docker and confirmed that the issue can also be resolved by patching our Cargo.toml, but I guess this is likely to be out of scope for is-docker since it adds support for non-Docker containers.

diff --git a/Cargo.toml b/Cargo.toml
index 1a7e482d66..0ee9faddd8 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -215,6 +215,9 @@ which = { version = "8.0.0" }
 wild = { version = "2" }
 zip = { version = "0.6.6", default-features = false }
 
+[patch.crates-io]
+is-docker = { path = "../other/is-docker" }
+
 [workspace.metadata.cargo-shear]
 ignored = [
     "getrandom",

diff --git a/src/lib.rs b/src/lib.rs
index 724f9f6..6d15416 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -3,13 +3,19 @@ extern crate once_cell;
 use std::fs;
 use once_cell::sync::OnceCell;
 
-fn has_docker_env_file() -> bool {
+fn has_env_file() -> bool {
     fs::metadata("/.dockerenv").is_ok()
+        || fs::metadata("/run/.containerenv").is_ok()
 }
 
 fn has_docker_in_cgroup() -> bool {
     match fs::read_to_string("/proc/self/cgroup") {
-        Ok(file_contents) => file_contents.contains("docker"),
+        Ok(file_contents) => {
+            file_contents.contains("docker")
+                || file_contents.contains("/container")
+                || file_contents.contains("/lxc")
+                || file_contents.contains("kubepods")
+        }
         Err(_error) => false,
     }
 }
@@ -18,6 +24,6 @@ pub fn is_docker() -> bool {
     static CACHED_RESULT: OnceCell<bool> = OnceCell::new();
 
     *CACHED_RESULT.get_or_init(|| {
-        has_docker_env_file() || has_docker_in_cgroup()
+        has_env_file() || has_docker_in_cgroup()
     })   
 }