Allow regex removal of non-utf8 env var names. (#149)

Fixes #148

CHANGES.md

@@ -1,5 +1,9 @@
 # Release Notes
 
+## 0.13.1
+
+Support regex removal of env vars with non-utf8 names in commands.
+
 ## 0.13.0
 
 This release improves the help screen for BusyBox scies with more clear messages for the various

Cargo.toml

@@ -5,7 +5,7 @@ members = [
 
 [package]
 name = "scie-jump"
-version = "0.13.0"
+version = "0.13.1"
 description = "The self contained interpreted executable launcher."
 authors = [
     "John Sirois <john.sirois@gmail.com>",

docs/packaging.md

@@ -152,7 +152,9 @@ variable names will be removed. For example, `"BASH_.*": null` would remove all
 variables whose name start with `BASH_` and `"=BASH_SOURCE": null` would just remove the
 `BASH_SOURCE` environment variable. When processing env entries, removals are done first, then
 defaults are set and finally overwrites are processed. This is regardless of the order of the env
-var entries in the lift manifest JSON document.
+var entries in the lift manifest JSON document. When evaluating environment variable removal regular
+expressions, the regular expression syntax is that supported by the Rust [`regex` crate](
+https://docs.rs/regex/latest/regex/).
 
 You can also supply a list of commands under "scie.lift.boot.bindings". These commands are objects
 with the same format as the "scie.lift.boot.commands" but they are not directly runnable by the end

jump/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "jump"
-version = "0.13.0"
+version = "0.13.1"
 description = "The bulk of the scie-jump binary logic."
 authors = [
     "John Sirois <john.sirois@gmail.com>",
@@ -21,7 +21,8 @@ itertools = "0.10"
 log = { workspace = true }
 logging_timer = { workspace = true }
 memmap2 = "0.7"
-regex = { version = "1.7", default-features = false, features = ["std"] }
+os_str_bytes = "6.5"
+regex = { version = "1.9", default-features = false, features = ["std"] }
 serde = { version = "1.0", features = ["derive"] }
 serde_json = "1.0"
 sha2 = "0.10"

jump/src/comparable_regex.rs

@@ -6,7 +6,7 @@ use std::fmt::{Display, Formatter};
 use std::hash::{Hash, Hasher};
 use std::ops::Deref;
 
-use regex::Regex;
+use regex::bytes::Regex;
 
 #[derive(Clone, Debug)]
 pub struct ComparableRegex(Regex);

jump/src/process.rs

@@ -7,6 +7,7 @@ use std::ffi::OsString;
 use std::process::{Child, Command, ExitStatus, Stdio};
 
 use logging_timer::time;
+use os_str_bytes::OsStrBytes;
 use sha2::{Digest, Sha256};
 
 use crate::comparable_regex::ComparableRegex;
@@ -64,19 +65,8 @@ impl EnvVars {
                 }
                 EnvVar::RemoveMatching(regex) => {
                     for (name, _) in env::vars_os() {
-                        match name.into_string() {
-                            Ok(name_utf8) => {
-                                if regex.is_match(&name_utf8) {
-                                    removals.insert(name_utf8.into());
-                                }
-                            }
-                            Err(name_os) => {
-                                warn!(
-                                    "Cannot process env removal matching regex '{regex}' for \
-                                    non-utf8 env var name {name_os:?}",
-                                    regex = regex.as_str()
-                                )
-                            }
+                        if regex.is_match(name.as_os_str().to_raw_bytes().as_ref()) {
+                            removals.insert(name);
                         }
                     }
                 }
@@ -213,6 +203,7 @@ mod tests {
     use std::ffi::OsString;
     use std::sync::{Arc, OnceLock};
 
+    use os_str_bytes::OsStrBytes;
     use parking_lot::ReentrantMutex;
 
     use crate::comparable_regex::ComparableRegex;
@@ -334,6 +325,25 @@ mod tests {
                 .to_env_vars(),
                 "Expected removal of an env var with a non-utf8 value to succeed."
             )
-        })
+        });
+
+        with_extra_env(&[(non_utf8.clone(), "baz".into())], || {
+            let mut re = String::new();
+            re.push('^');
+            for byte in non_utf8.as_os_str().to_raw_bytes().as_ref() {
+                re.push_str(format!(r"(?-u:\x{:X})", byte).as_str());
+            }
+            re.push('$');
+            assert_eq!(
+                vec![(non_utf8.clone(), None)],
+                EnvVars {
+                    vars: vec![EnvVar::RemoveMatching(
+                        ComparableRegex::try_from(re.as_str()).unwrap()
+                    )]
+                }
+                .to_env_vars(),
+                "Expected removal of an env var with a non-utf8 name to succeed."
+            )
+        });
     }
 }