fix(no-proxy-pattern): stop chr() corrupting IP bytes as lossy UTF-8

shirabe_php_shim::chr() returned a Rust String, which lossily
re-encodes bytes >= 0x80 as UTF-8 replacement characters. ip_get_mask,
ip_get_network, and ip_map_to_6 relied on chr() to build raw in_addr
and netmask byte arrays, corrupting IPv4-in-IPv6 mappings and CIDR
netmasks. Build the Vec<u8> byte arrays directly instead of
round-tripping through String, and un-ignore test_ip_address and
test_ip_range now that the underlying bug is fixed.

chr()'s only other caller (http_downloader.rs, an ASCII ESC byte in a
regex pattern) didn't need the indirection either, so remove the shim
function entirely.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

4 files changed, 16 insertions, 32 deletions

diff --git a/crates/shirabe-php-shim/src/string.rs b/crates/shirabe-php-shim/src/string.rs
index 97c66953..2617399f 100644
--- a/crates/shirabe-php-shim/src/string.rs
+++ b/crates/shirabe-php-shim/src/string.rs
@@ -1289,12 +1289,6 @@ pub fn ucfirst(s: &str) -> String {
     }
 }
 
-pub fn chr(_value: u8) -> String {
-    // TODO(phase-d): PHP chr() yields a single raw byte; for values >= 0x80 that byte is not valid
-    // UTF-8, so storing it in a Rust String forces a lossy substitution.
-    String::from_utf8_lossy(&[_value]).into_owned()
-}
-
 // Port of PHP's addcslashes: every byte that falls in the (range-expanded) charlist is
 // backslash-escaped, with non-printable bytes rendered as the C escape or a three-digit octal.
 pub fn addcslashes(_string: &str, _charlist: &str) -> String {
diff --git a/crates/shirabe/src/util/http_downloader.rs b/crates/shirabe/src/util/http_downloader.rs
index b055a515..0c91484d 100644
--- a/crates/shirabe/src/util/http_downloader.rs
+++ b/crates/shirabe/src/util/http_downloader.rs
@@ -18,9 +18,9 @@ use crate::util::sync_executor;
 use indexmap::IndexMap;
 use shirabe_external_packages::composer::pcre::{CaptureKey, Preg};
 use shirabe_php_shim::{
-    InvalidArgumentException, LogicException, PhpMixed, array_replace_recursive, chr,
-    extension_loaded, file_get_contents, function_exists, implode, is_numeric, php_regex,
-    rawurldecode, stream_context_create, stripos, strpos, substr, ucfirst,
+    InvalidArgumentException, LogicException, PhpMixed, array_replace_recursive, extension_loaded,
+    file_get_contents, function_exists, implode, is_numeric, php_regex, rawurldecode,
+    stream_context_create, stripos, strpos, substr, ucfirst,
 };
 use shirabe_semver::constraint::SimpleConstraint;
 
@@ -381,11 +381,7 @@ impl HttpDownloader {
     ) -> anyhow::Result<()> {
         let clean_message = |msg: &str| -> anyhow::Result<String> {
             if !io.is_decorated() {
-                return Ok(Preg::replace(
-                    format!("{{{}{}}}u", chr(27), "\\[[;\\d]*m"),
-                    "",
-                    msg,
-                ));
+                return Ok(Preg::replace("{\x1b\\[[;\\d]*m}u", "", msg));
             }
 
             Ok(msg.to_string())
diff --git a/crates/shirabe/src/util/no_proxy_pattern.rs b/crates/shirabe/src/util/no_proxy_pattern.rs
index dc91a256..a33f2451 100644
--- a/crates/shirabe/src/util/no_proxy_pattern.rs
+++ b/crates/shirabe/src/util/no_proxy_pattern.rs
@@ -3,9 +3,9 @@
 use indexmap::IndexMap;
 use shirabe_external_packages::composer::pcre::Preg;
 use shirabe_php_shim::{
-    PHP_URL_HOST, PHP_URL_PORT, PHP_URL_SCHEME, PhpMixed, RuntimeException, array_key_exists, chr,
+    PHP_URL_HOST, PHP_URL_PORT, PHP_URL_SCHEME, PhpMixed, RuntimeException, array_key_exists,
     empty, explode, filter_var_int_with_range, filter_var_ip, inet_pton, ltrim, parse_url,
-    php_regex, str_pad, str_repeat, stripos, strlen, strpbrk, strpos, substr, substr_count, unpack,
+    php_regex, stripos, strlen, strpbrk, strpos, substr, substr_count, unpack,
 };
 
 /// Tests URLs against NO_PROXY patterns
@@ -315,26 +315,23 @@ impl NoProxyPattern {
     /// @param int $prefix CIDR prefix-length
     /// @param int $size   Byte size of in_addr
     fn ip_get_mask(&self, prefix: i64, size: i64) -> Vec<u8> {
-        let mut mask = String::new();
+        let mut mask: Vec<u8> = Vec::new();
 
         let ones = (prefix as f64 / 8.0).floor() as i64;
         if ones != 0 {
-            mask = str_repeat(&chr(255), ones as usize);
+            mask.extend(vec![0xffu8; ones as usize]);
         }
 
         let remainder = prefix % 8;
         if remainder != 0 {
-            mask.push_str(&chr(0xff ^ (0xff >> remainder)));
+            mask.push(0xff ^ (0xff >> remainder));
         }
 
-        let mask = str_pad(
-            &mask,
-            size as usize,
-            &chr(0),
-            shirabe_php_shim::STR_PAD_RIGHT,
-        );
+        if (size as usize) > mask.len() {
+            mask.resize(size as usize, 0);
+        }
 
-        self.ip_map_to_6(mask.as_bytes(), size)
+        self.ip_map_to_6(&mask, size)
     }
 
     /// Calculates and returns the network and mask
@@ -389,8 +386,7 @@ impl NoProxyPattern {
                 .get(&i.to_string())
                 .and_then(|v| v.as_int())
                 .unwrap_or(0);
-            // PHP: $net .= chr($ip[$i] & $mask[$i]);
-            net.extend(chr((ip_byte & mask_byte) as u8).as_bytes());
+            net.push((ip_byte & mask_byte) as u8);
         }
 
         Ok((net, netmask))
@@ -404,8 +400,8 @@ impl NoProxyPattern {
     /// @return string Mapped or existing in_addr
     fn ip_map_to_6(&self, binary: &[u8], size: i64) -> Vec<u8> {
         if size == 4 {
-            let mut prefix = str_repeat(&chr(0), 10).into_bytes();
-            prefix.extend(str_repeat(&chr(255), 2).into_bytes());
+            let mut prefix = vec![0u8; 10];
+            prefix.extend(vec![0xffu8; 2]);
             prefix.extend_from_slice(binary);
             return prefix;
         }
diff --git a/crates/shirabe/tests/util/no_proxy_pattern_test.rs b/crates/shirabe/tests/util/no_proxy_pattern_test.rs
index e59e28a6..63f0d5f6 100644
--- a/crates/shirabe/tests/util/no_proxy_pattern_test.rs
+++ b/crates/shirabe/tests/util/no_proxy_pattern_test.rs
@@ -42,7 +42,6 @@ fn test_host_name() {
 }
 
 #[test]
-#[ignore = "shirabe_php_shim::chr() lossily re-encodes bytes >= 0x80 as UTF-8 (TODO(phase-d) in its own doc comment); ip_map_to_6's chr(255) filler bytes become 3-byte U+FFFD replacements, corrupting the mapped IPv4-in-IPv6 byte arrays used for comparison"]
 fn test_ip_address() {
     let noproxy = "192.168.1.1, 2001:db8::52:0:1";
 
@@ -55,7 +54,6 @@ fn test_ip_address() {
 }
 
 #[test]
-#[ignore = "shirabe_php_shim::chr() lossily re-encodes bytes >= 0x80 as UTF-8 (TODO(phase-d) in its own doc comment); ip_get_mask's chr(255)/chr(0xff^...) calls corrupt the CIDR netmask bytes used in match_range, breaking all prefix-based no_proxy matching"]
 fn test_ip_range() {
     let noproxy = "10.0.0.0/30, 2002:db8:a::45/121";