From 880ba0fd1d05faf98588cc326b3d9fbe625ebf2b Mon Sep 17 00:00:00 2001
From: nsfisis <nsfisis@gmail.com>
Date: Sun, 9 Aug 2026 11:13:06 +0900
Subject: refactor(symfony-string): extract symfony/string into the
 shirabe-symfony-string crate

Move `Symfony\Component\String` out of shirabe-external-packages and into
its own crate, so the path is `shirabe_symfony_string::ByteString`
instead of `shirabe_external_packages::symfony::string::ByteString`.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---
 .../src/code_point_string.rs                       | 93 ++++++++++++++++++++++
 1 file changed, 93 insertions(+)
 create mode 100644 crates/shirabe-symfony-string/src/code_point_string.rs

(limited to 'crates/shirabe-symfony-string/src/code_point_string.rs')

diff --git a/crates/shirabe-symfony-string/src/code_point_string.rs b/crates/shirabe-symfony-string/src/code_point_string.rs
new file mode 100644
index 00000000..dcbaa1f6
--- /dev/null
+++ b/crates/shirabe-symfony-string/src/code_point_string.rs
@@ -0,0 +1,93 @@
+//! ref: composer/vendor/symfony/string/CodePointString.php
+
+#[derive(Debug, Clone)]
+pub struct CodePointString {
+    pub(crate) string: String,
+}
+
+impl CodePointString {
+    /// Port of `AbstractString::wordwrap()`, specialised to the non-`ignoreCase` code-point case.
+    pub fn wordwrap(&self, width: i64, r#break: &str, cut: bool) -> Self {
+        // `split($break)` with no flags reduces to `explode($break, $string)` here, then `chunk()`
+        // yields one entry per code point. `ignoreCase` is always false for freshly built instances.
+        let lines: Vec<&str> = if !r#break.is_empty() {
+            self.string.split(r#break).collect()
+        } else {
+            vec![&self.string]
+        };
+
+        let mut chars: Vec<String> = Vec::new();
+        let mut mask = String::new();
+
+        if lines.len() == 1 && lines[0].is_empty() {
+            return Self {
+                string: String::new(),
+            };
+        }
+
+        for (i, line) in lines.iter().enumerate() {
+            if i != 0 {
+                chars.push(r#break.to_string());
+                mask.push('#');
+            }
+
+            for ch in line.chars() {
+                let s = ch.to_string();
+                mask.push(if s == " " { ' ' } else { '?' });
+                chars.push(s);
+            }
+        }
+
+        let mut string = String::new();
+        let mut j: usize = 0;
+        // PHP seeds both `$b` and `$i` at -1; mirror with signed indices.
+        let mut i: i64 = -1;
+        let mask = shirabe_php_shim::wordwrap(&mask, width, "#", cut);
+        let mask_bytes = mask.as_bytes();
+
+        let mut b: i64 = -1;
+        loop {
+            // strpos($mask, '#', $b + 1)
+            let from = (b + 1) as usize;
+            let Some(rel) = mask_bytes[from..].iter().position(|&c| c == b'#') else {
+                break;
+            };
+            b = (from + rel) as i64;
+
+            i += 1;
+            while i < b {
+                string.push_str(&chars[j]);
+                j += 1;
+                i += 1;
+            }
+
+            if chars[j] == r#break || chars[j] == " " {
+                j += 1;
+            }
+
+            string.push_str(r#break);
+        }
+
+        for c in &chars[j..] {
+            string.push_str(c);
+        }
+
+        Self { string }
+    }
+
+    /// `to_encoding` is `""` for PHP's `null`.
+    pub fn to_byte_string(&self, to_encoding: &str) -> String {
+        // A CodePointString is an AbstractUnicodeString, so PHP's `$fromEncoding` is always
+        // 'UTF-8' and the string is returned verbatim for a null/UTF-8 target.
+        if matches!(to_encoding, "" | "utf8" | "utf-8" | "UTF8" | "UTF-8") {
+            return self.string.clone();
+        }
+
+        // PHP falls back to iconv() only when mb_convert_encoding() rejects the target encoding.
+        shirabe_php_shim::mb_convert_encoding(
+            self.string.clone().into_bytes(),
+            to_encoding,
+            "UTF-8",
+        )
+    }
+}
-- 
cgit v1.3.1-4-g156e