From f5787184dae2e98a736f63225aafe90a68923514 Mon Sep 17 00:00:00 2001
From: Justin Ethier <justin.ethier@gmail.com>
Date: Tue, 7 Nov 2017 18:18:56 +0000
Subject: [PATCH] WIP - string-set!

---
 runtime.c | 22 ++++++++++++++++++----
 1 file changed, 18 insertions(+), 4 deletions(-)

diff --git a/runtime.c b/runtime.c
index e7792702..f129899a 100644
--- a/runtime.c
+++ b/runtime.c
@@ -2117,8 +2117,10 @@ object Cyc_string_byte_length(void *data, object str)
 
 object Cyc_string_set(void *data, object str, object k, object chr)
 {
+  char buf[5];
   char *raw;
-  int idx, len;
+  int idx, len, buf_len;
+  char_type input_char;
 
   Cyc_check_str(data, str);
   Cyc_check_num(data, k);
@@ -2127,6 +2129,10 @@ object Cyc_string_set(void *data, object str, object k, object chr)
     Cyc_rt_raise2(data, "Expected char but received", chr);
   }
 
+  input_char = obj_obj2char(chr);
+  Cyc_utf8_encode_char(buf, 5, input_char);
+  buf_len = strlen(buf);
+
   raw = string_str(str);
   idx = unbox_number(k);
   len = string_len(str);
@@ -2134,8 +2140,7 @@ object Cyc_string_set(void *data, object str, object k, object chr)
   Cyc_check_bounds(data, "string-set!", len, idx);
 
   // Take fast path if all chars are just 1 byte
-  if (string_num_cp(str) == string_len(str)) {
-    // TODO: not good enough, chr could be multi-byte
+  if (string_num_cp(str) == string_len(str) && buf_len == 1) {
     raw[idx] = obj_obj2char(chr);
   } else {
 fprintf(stderr, "DEBUG %s, num_cp = %d, len = %d\n", raw, string_num_cp(str), len);
@@ -2148,9 +2153,10 @@ fprintf(stderr, "DEBUG %s, num_cp = %d, len = %d\n", raw, string_num_cp(str), le
     char *tmp = raw;
     char_type codepoint;
     uint32_t state = 0;
-    int i = 0, count, start_len = 0, start_cp = 0;
+    int i = 0, count, start_len = 0, start_cp = 0, bytes = 0;
 
     for (count = 0; *tmp; ++tmp){
+      bytes++;
       if (!Cyc_utf8_decode(&state, &codepoint, (uint8_t)*tmp)){
         if (count < idx) {
           start_len = i;
@@ -2159,6 +2165,7 @@ fprintf(stderr, "DEBUG %s, num_cp = %d, len = %d\n", raw, string_num_cp(str), le
           break;
         }
         count += 1;
+        bytes = 0;
       }
       i++;
     }
@@ -2171,6 +2178,13 @@ fprintf(stderr, "DEBUG %s, num_cp = %d, len = %d\n", raw, string_num_cp(str), le
     // and we know the codepoint to be replaced. by calculating its length
     // we can compute where the end portion starts, and by using str we can
     // figure out how many remaining bytes/codepoints are in end
+    //
+    // 3 cases: 
+    // - buf_len = bytes, just straight replace
+    // - buf_len > bytes, will need to allocate more memory (!!)
+    // - buf_len < bytes, just replace, but pad with NULL chars.
+    //                    in this case need to ensure string_len is not 
+    //                    reduced because original value still matters for GC purposes
 
   }
   return str;