From 0f0fc2db3549c599063c9dfc0dbde76cda238eff Mon Sep 17 00:00:00 2001
From: Blendi-Goose <87442375+Blendi-Goose@users.noreply.github.com>
Date: Sun, 8 Jun 2025 23:56:45 +0200
Subject: [PATCH] absolutely awful progress that needs to be checked by every
 developer before it can possibly ever be trusted because i wrote it while
 very much not properly awake, and i'm not awake now to fix it either.

---
 src/unicode.c | 45 ++++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 40 insertions(+), 5 deletions(-)

diff --git a/src/unicode.c b/src/unicode.c
index a4cf285..6912f1b 100644
--- a/src/unicode.c
+++ b/src/unicode.c
@@ -12,17 +12,17 @@ bool nn_unicode_validate(const char *s) {
 
 char *nn_unicode_char(int *codepoints, size_t codepointCount) {
     size_t len = 0;
-    for(size_t i = 0; i < codepointCount; i++) {
+    for (size_t i = 0; i < codepointCount; i++) {
         int codepoint = codepoints[i];
         len += nn_unicode_codepointSize(codepoint);
     }
 
     char *buf = nn_malloc(len+1);
-    if(buf == NULL) return buf;
+    if (buf == NULL) return buf;
     buf[len] = '\0';
 
     size_t j = 0;
-    for(size_t i = 0; i < codepointCount; i++) {
+    for (size_t i = 0; i < codepointCount; i++) {
         int codepoint = codepoints[i];
         size_t codepointLen = 0;
         const char *c = nn_unicode_codepointToChar(codepoint, &codepointLen);
@@ -45,9 +45,44 @@ size_t nn_unicode_len(const char *s) {
 
 int nn_unicode_codepointAt(const char *s, size_t byteOffset);
 
-size_t nn_unicode_codepointSize(int codepoint);
+size_t nn_unicode_codepointSize(int codepoint) {
+    if (codepoint <= 0x007f) {
+        return 1;
+    } else if (codepoint <= 0x07ff) {
+        return 2;
+    } else if (codepoint <= 0xffff) {
+        return 3;
+    } else if (codepoint <= 0x10ffff) {
+        return 4;
+    }
 
-const char *nn_unicode_codepointToChar(int codepoint, size_t *len);
+    return 0;
+}
+
+const char *nn_unicode_codepointToChar(int codepoint, size_t *len) {
+    size_t codepointSize = nn_unicode_codepointSize(codepoint);
+    *len = codepointSize;
+
+    static char buffer[4];
+
+    if (codepointSize == 1) {
+        buffer[0] = (char)codepoint;
+    } else if (codepointSize == 2) {
+        buffer[0] = 0b11000000 + (codepoint & 0b11111);
+        buffer[1] = 0b10000000 + (codepoint >> 5);
+    } else if (codepointSize == 3) {
+        buffer[0] = 0b11100000 + (codepoint & 0b1111);
+        buffer[1] = 0b10000000 + ((codepoint >> 4) & 0b111111);
+        buffer[2] = 0b10000000 + (codepoint >> 10);
+    } else if (codepointSize == 4) {
+        buffer[0] = 0b11110000 + (codepoint & 0b111);
+        buffer[1] = 0b10000000 + ((codepoint >> 3) & 0b111111);
+        buffer[2] = 0b10000000 + ((codepoint >> 9) & 0b111111);
+        buffer[3] = 0b10000000 + (codepoint >> 15);
+    }
+
+    return buffer;
+}
 
 size_t nn_unicode_charWidth(int codepoint);