From ec563d33a05580eff1b2561e512a46b5b0898d00 Mon Sep 17 00:00:00 2001
From: Adam Cimarosti <cimarosti@gmail.com>
Date: Mon, 29 Jan 2024 12:28:50 +0000
Subject: [PATCH] feat: excluded control chars in values

---
 README.md      |  7 ++++---
 src/lib.rs     |  7 ++++++-
 tests/tests.rs | 19 +++++++++++++++++++
 3 files changed, 29 insertions(+), 4 deletions(-)

diff --git a/README.md b/README.md
index b049537..2a89f85 100644
--- a/README.md
+++ b/README.md
@@ -14,9 +14,10 @@ service::key1=value1;key2=value2;key3=value3;
 A few rules:
 * The last semicolon is mandatory.
 * Service name and keys are case-insensitive.
-* Values are case-sensitive.
 * Keys are ASCII alphanumeric and start with a letter.
-* If a semicolon `;` appears in a value, escaped it as a double semicolon `;;`.
+* Values are case-sensitive unicode strings which can contain any characters,
+  * Except control characters (`0x00..=0x1f` and `0x7f..=0x9f`).
+  * If semicolons `;` appears in a value, these are escaped as double semicolon `;;`.
 
 ## Grammar
 
@@ -33,7 +34,7 @@ alpha ::= "a".."z" | "A".."Z"
 alphanumeric ::= "a".."z" | "A".."Z" | "0".."9"
 value_char ::= non_semicolon_char | escaped_semicolon
 escaped_semicolon ::= ";;"
-non_semicolon_char ::= ? any character except ';' ?
+non_semicolon_char ::= ? any unicode character except ';', 0x00..=0x1f and 0x7f..=0x9f ?
 ```
 
 ## Usage
diff --git a/src/lib.rs b/src/lib.rs
index 1d4ad95..bcb2baf 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -91,6 +91,7 @@ pub enum ErrorKind {
     ExpectedIdentifierNotEmpty,
     BadSeparator((char, char)),
     IncompleteKeyValue,
+    InvalidValueChar(char),
     MissingTrailingSemicolon,
     DuplicateKey(String),
 }
@@ -128,6 +129,7 @@ impl Display for ErrorKind {
             ErrorKind::IncompleteKeyValue => {
                 write!(f, "incomplete key-value pair before end of input")
             }
+            ErrorKind::InvalidValueChar(c) => write!(f, "invalid value char {:?}", c),
             ErrorKind::MissingTrailingSemicolon => write!(f, "missing trailing semicolon"),
             ErrorKind::DuplicateKey(s) => write!(f, "duplicate key {:?}", s),
         }
@@ -215,7 +217,10 @@ fn parse_value(
                 value.push(';');
             }
             (Some((_, ';')), _) => break,
-            (Some((_, c)), _) => {
+            (Some((p, c)), _) => {
+                if matches!(c, '\u{0}'..='\u{1f}' | '\u{7f}'..='\u{9f}') {
+                    return Err(parse_err(ErrorKind::InvalidValueChar(c), p));
+                }
                 value.push(c);
                 let _ = iter.next();
             }
diff --git a/tests/tests.rs b/tests/tests.rs
index f3988a8..f537351 100644
--- a/tests/tests.rs
+++ b/tests/tests.rs
@@ -328,3 +328,22 @@ fn unicode_value() -> Result<(), ParsingError> {
     assert_eq!(config.get("x"), Some("協定"));
     Ok(())
 }
+
+#[test]
+fn invalid_ctrl_chars_in_value() {
+    let bad_chars = [
+        '\x00', '\x01', '\x02', '\x03', '\x04', '\x1f', '\x7f', '\u{80}', '\u{8a}', '\u{9f}',
+    ];
+    for bad in bad_chars {
+        let input = format!("http::x={};", bad);
+        let config = parse_conf_str(&input);
+        assert!(config.is_err());
+        let err = config.unwrap_err();
+        assert_eq!(err.kind(), ErrorKind::InvalidValueChar(bad));
+        assert_eq!(err.position(), 8);
+        assert_eq!(
+            err.to_string(),
+            format!("invalid value char {:?} at position 8", bad)
+        );
+    }
+}