WARNING: Several changes on UTF-8/Codepoints API

 - ADDED: `GetCodepointPrevious()`
 - RENAMED: `GetCodepoint()` -> `GetCodepointNext()`, actually, reimplemented
 - `GetCodepoint()` has been kept for the moment, for compatibility and also because implementation is different
 - RENAMED: `TextCodepointsToUTF8()` to `LoadUTF8()`, simpler name and more aligned with raylib conventions (functions loading memory start with Load*()), parameters should be descriptive of functionailty.
 - ADDED: `UnloadUTF8()`, aligned with `LoadUTF8()` to avoid allocators issues.

3 files changed, 136 insertions, 72 deletions

diff --git a/src/raylib.h b/src/raylib.h
index 7e4fdf3f..84a18ebe 100644
--- a/src/raylib.h
+++ b/src/raylib.h
@@ -1370,12 +1370,15 @@ RLAPI GlyphInfo GetGlyphInfo(Font font, int codepoint);
 RLAPI Rectangle GetGlyphAtlasRec(Font font, int codepoint);                                 // Get glyph rectangle in font atlas for a codepoint (unicode character), fallback to '?' if not found
 
 // Text codepoints management functions (unicode characters)
-RLAPI int *LoadCodepoints(const char *text, int *count);              // Load all codepoints from a UTF-8 text string, codepoints count returned by parameter
-RLAPI void UnloadCodepoints(int *codepoints);                         // Unload codepoints data from memory
-RLAPI int GetCodepointCount(const char *text);                        // Get total number of codepoints in a UTF-8 encoded string
-RLAPI int GetCodepoint(const char *text, int *bytesProcessed);        // Get next codepoint in a UTF-8 encoded string, 0x3f('?') is returned on failure
-RLAPI const char *CodepointToUTF8(int codepoint, int *byteSize);      // Encode one codepoint into UTF-8 byte array (array length returned as parameter)
-RLAPI char *TextCodepointsToUTF8(const int *codepoints, int length);  // Encode text as codepoints array into UTF-8 text string (WARNING: memory must be freed!)
+RLAPI char *LoadUTF8(const int *codepoints, int length);                // Load UTF-8 text encoded from codepoints array
+RLAPI void UnloadUTF8(char *text);                                      // Unload UTF-8 text encoded from codepoints array
+RLAPI int *LoadCodepoints(const char *text, int *count);                // Load all codepoints from a UTF-8 text string, codepoints count returned by parameter
+RLAPI void UnloadCodepoints(int *codepoints);                           // Unload codepoints data from memory
+RLAPI int GetCodepointCount(const char *text);                          // Get total number of codepoints in a UTF-8 encoded string
+RLAPI int GetCodepoint(const char *text, int *codepointSize);           // Get next codepoint in a UTF-8 encoded string, 0x3f('?') is returned on failure
+RLAPI int GetCodepointNext(const char *text, int *codepointSize);       // Get next codepoint in a UTF-8 encoded string, 0x3f('?') is returned on failure
+RLAPI int GetCodepointPrevious(const char *text, int *codepointSize);   // Get previous codepoint in a UTF-8 encoded string, 0x3f('?') is returned on failure
+RLAPI const char *CodepointToUTF8(int codepoint, int *utf8Size);        // Encode one codepoint into UTF-8 byte array (array length returned as parameter)
 
 // Text strings management functions (no UTF-8 strings, only byte chars)
 // NOTE: Some strings allocate memory internally for returned strings, just be careful!
diff --git a/src/rtext.c b/src/rtext.c
index ef5debd5..324ebe95 100644
--- a/src/rtext.c
+++ b/src/rtext.c
@@ -1043,7 +1043,7 @@ void DrawTextEx(Font font, const char *text, Vector2 position, float fontSize, f
     {
         // Get next codepoint from byte string and glyph index in font
         int codepointByteCount = 0;
-        int codepoint = GetCodepoint(&text[i], &codepointByteCount);
+        int codepoint = GetCodepointNext(&text[i], &codepointByteCount);
         int index = GetGlyphIndex(font, codepoint);
 
         // NOTE: Normally we exit the decoding sequence as soon as a bad byte is found (and return 0x3f)
@@ -1185,7 +1185,7 @@ Vector2 MeasureTextEx(Font font, const char *text, float fontSize, float spacing
         byteCounter++;
 
         int next = 0;
-        letter = GetCodepoint(&text[i], &next);
+        letter = GetCodepointNext(&text[i], &next);
         index = GetGlyphIndex(font, letter);
 
         // NOTE: normally we exit the decoding sequence as soon as a bad byte is found (and return 0x3f)
@@ -1627,7 +1627,7 @@ const char *TextToPascal(const char *text)
 // Encode text codepoint into UTF-8 text
 // REQUIRES: memcpy()
 // WARNING: Allocated memory must be manually freed
-char *TextCodepointsToUTF8(const int *codepoints, int length)
+char *LoadUTF8(const int *codepoints, int length)
 {
     // We allocate enough memory fo fit all possible codepoints
     // NOTE: 5 bytes for every codepoint should be enough
@@ -1650,43 +1650,10 @@ char *TextCodepointsToUTF8(const int *codepoints, int length)
     return text;
 }
 
-// Encode codepoint into utf8 text (char array length returned as parameter)
-// NOTE: It uses a static array to store UTF-8 bytes
-RLAPI const char *CodepointToUTF8(int codepoint, int *byteSize)
+// Unload UTF-8 text encoded from codepoints array
+void UnloadUTF8(char *text)
 {
-    static char utf8[6] = { 0 };
-    int size = 0;   // Byte size of codepoint
-
-    if (codepoint <= 0x7f)
-    {
-        utf8[0] = (char)codepoint;
-        size = 1;
-    }
-    else if (codepoint <= 0x7ff)
-    {
-        utf8[0] = (char)(((codepoint >> 6) & 0x1f) | 0xc0);
-        utf8[1] = (char)((codepoint & 0x3f) | 0x80);
-        size = 2;
-    }
-    else if (codepoint <= 0xffff)
-    {
-        utf8[0] = (char)(((codepoint >> 12) & 0x0f) | 0xe0);
-        utf8[1] = (char)(((codepoint >>  6) & 0x3f) | 0x80);
-        utf8[2] = (char)((codepoint & 0x3f) | 0x80);
-        size = 3;
-    }
-    else if (codepoint <= 0x10ffff)
-    {
-        utf8[0] = (char)(((codepoint >> 18) & 0x07) | 0xf0);
-        utf8[1] = (char)(((codepoint >> 12) & 0x3f) | 0x80);
-        utf8[2] = (char)(((codepoint >>  6) & 0x3f) | 0x80);
-        utf8[3] = (char)((codepoint & 0x3f) | 0x80);
-        size = 4;
-    }
-
-    *byteSize = size;
-
-    return utf8;
+    RL_FREE(text);
 }
 
 // Load all codepoints from a UTF-8 text string, codepoints count returned by parameter
@@ -1694,7 +1661,7 @@ int *LoadCodepoints(const char *text, int *count)
 {
     int textLength = TextLength(text);
 
-    int bytesProcessed = 0;
+    int codepointSize = 0;
     int codepointCount = 0;
 
     // Allocate a big enough buffer to store as many codepoints as text bytes
@@ -1702,8 +1669,8 @@ int *LoadCodepoints(const char *text, int *count)
 
     for (int i = 0; i < textLength; codepointCount++)
     {
-        codepoints[codepointCount] = GetCodepoint(text + i, &bytesProcessed);
-        i += bytesProcessed;
+        codepoints[codepointCount] = GetCodepointNext(text + i, &codepointSize);
+        i += codepointSize;
     }
 
     // Re-allocate buffer to the actual number of codepoints loaded
@@ -1731,7 +1698,7 @@ int GetCodepointCount(const char *text)
     while (*ptr != '\0')
     {
         int next = 0;
-        int letter = GetCodepoint(ptr, &next);
+        int letter = GetCodepointNext(ptr, &next);
 
         if (letter == 0x3f) ptr += 1;
         else ptr += next;
@@ -1741,6 +1708,45 @@ int GetCodepointCount(const char *text)
 
     return length;
 }
+
+// Encode codepoint into utf8 text (char array length returned as parameter)
+// NOTE: It uses a static array to store UTF-8 bytes
+const char *CodepointToUTF8(int codepoint, int *utf8Size)
+{
+    static char utf8[6] = { 0 };
+    int size = 0;   // Byte size of codepoint
+
+    if (codepoint <= 0x7f)
+    {
+        utf8[0] = (char)codepoint;
+        size = 1;
+    }
+    else if (codepoint <= 0x7ff)
+    {
+        utf8[0] = (char)(((codepoint >> 6) & 0x1f) | 0xc0);
+        utf8[1] = (char)((codepoint & 0x3f) | 0x80);
+        size = 2;
+    }
+    else if (codepoint <= 0xffff)
+    {
+        utf8[0] = (char)(((codepoint >> 12) & 0x0f) | 0xe0);
+        utf8[1] = (char)(((codepoint >>  6) & 0x3f) | 0x80);
+        utf8[2] = (char)((codepoint & 0x3f) | 0x80);
+        size = 3;
+    }
+    else if (codepoint <= 0x10ffff)
+    {
+        utf8[0] = (char)(((codepoint >> 18) & 0x07) | 0xf0);
+        utf8[1] = (char)(((codepoint >> 12) & 0x3f) | 0x80);
+        utf8[2] = (char)(((codepoint >>  6) & 0x3f) | 0x80);
+        utf8[3] = (char)((codepoint & 0x3f) | 0x80);
+        size = 4;
+    }
+
+    *utf8Size = size;
+
+    return utf8;
+}
 #endif      // SUPPORT_TEXT_MANIPULATION
 
 // Get next codepoint in a UTF-8 encoded text, scanning until '\0' is found
@@ -1748,7 +1754,7 @@ int GetCodepointCount(const char *text)
 // Total number of bytes processed are returned as a parameter
 // NOTE: The standard says U+FFFD should be returned in case of errors
 // but that character is not supported by the default font in raylib
-int GetCodepoint(const char *text, int *bytesProcessed)
+int GetCodepoint(const char *text, int *codepointSize)
 {
 /*
     UTF-8 specs from https://www.ietf.org/rfc/rfc3629.txt
@@ -1763,14 +1769,14 @@ int GetCodepoint(const char *text, int *bytesProcessed)
 */
     // NOTE: on decode errors we return as soon as possible
 
-    int code = 0x3f;   // Codepoint (defaults to '?')
+    int codepoint = 0x3f;   // Codepoint (defaults to '?')
     int octet = (unsigned char)(text[0]); // The first UTF8 octet
-    *bytesProcessed = 1;
+    *codepointSize = 1;
 
     if (octet <= 0x7f)
     {
         // Only one octet (ASCII range x00-7F)
-        code = text[0];
+        codepoint = text[0];
     }
     else if ((octet & 0xe0) == 0xc0)
     {
@@ -1779,12 +1785,12 @@ int GetCodepoint(const char *text, int *bytesProcessed)
         // [0]xC2-DF    [1]UTF8-tail(x80-BF)
         unsigned char octet1 = text[1];
 
-        if ((octet1 == '\0') || ((octet1 >> 6) != 2)) { *bytesProcessed = 2; return code; } // Unexpected sequence
+        if ((octet1 == '\0') || ((octet1 >> 6) != 2)) { *codepointSize = 2; return codepoint; } // Unexpected sequence
 
         if ((octet >= 0xc2) && (octet <= 0xdf))
         {
-            code = ((octet & 0x1f) << 6) | (octet1 & 0x3f);
-            *bytesProcessed = 2;
+            codepoint = ((octet & 0x1f) << 6) | (octet1 & 0x3f);
+            *codepointSize = 2;
         }
     }
     else if ((octet & 0xf0) == 0xe0)
@@ -1793,11 +1799,11 @@ int GetCodepoint(const char *text, int *bytesProcessed)
         unsigned char octet1 = text[1];
         unsigned char octet2 = '\0';
 
-        if ((octet1 == '\0') || ((octet1 >> 6) != 2)) { *bytesProcessed = 2; return code; } // Unexpected sequence
+        if ((octet1 == '\0') || ((octet1 >> 6) != 2)) { *codepointSize = 2; return codepoint; } // Unexpected sequence
 
         octet2 = text[2];
 
-        if ((octet2 == '\0') || ((octet2 >> 6) != 2)) { *bytesProcessed = 3; return code; } // Unexpected sequence
+        if ((octet2 == '\0') || ((octet2 >> 6) != 2)) { *codepointSize = 3; return codepoint; } // Unexpected sequence
 
         // [0]xE0    [1]xA0-BF       [2]UTF8-tail(x80-BF)
         // [0]xE1-EC [1]UTF8-tail    [2]UTF8-tail(x80-BF)
@@ -1805,50 +1811,105 @@ int GetCodepoint(const char *text, int *bytesProcessed)
         // [0]xEE-EF [1]UTF8-tail    [2]UTF8-tail(x80-BF)
 
         if (((octet == 0xe0) && !((octet1 >= 0xa0) && (octet1 <= 0xbf))) ||
-            ((octet == 0xed) && !((octet1 >= 0x80) && (octet1 <= 0x9f)))) { *bytesProcessed = 2; return code; }
+            ((octet == 0xed) && !((octet1 >= 0x80) && (octet1 <= 0x9f)))) { *codepointSize = 2; return codepoint; }
 
         if ((octet >= 0xe0) && (octet <= 0xef))
         {
-            code = ((octet & 0xf) << 12) | ((octet1 & 0x3f) << 6) | (octet2 & 0x3f);
-            *bytesProcessed = 3;
+            codepoint = ((octet & 0xf) << 12) | ((octet1 & 0x3f) << 6) | (octet2 & 0x3f);
+            *codepointSize = 3;
         }
     }
     else if ((octet & 0xf8) == 0xf0)
     {
         // Four octets
-        if (octet > 0xf4) return code;
+        if (octet > 0xf4) return codepoint;
 
         unsigned char octet1 = text[1];
         unsigned char octet2 = '\0';
         unsigned char octet3 = '\0';
 
-        if ((octet1 == '\0') || ((octet1 >> 6) != 2)) { *bytesProcessed = 2; return code; }  // Unexpected sequence
+        if ((octet1 == '\0') || ((octet1 >> 6) != 2)) { *codepointSize = 2; return codepoint; }  // Unexpected sequence
 
         octet2 = text[2];
 
-        if ((octet2 == '\0') || ((octet2 >> 6) != 2)) { *bytesProcessed = 3; return code; }  // Unexpected sequence
+        if ((octet2 == '\0') || ((octet2 >> 6) != 2)) { *codepointSize = 3; return codepoint; }  // Unexpected sequence
 
         octet3 = text[3];
 
-        if ((octet3 == '\0') || ((octet3 >> 6) != 2)) { *bytesProcessed = 4; return code; }  // Unexpected sequence
+        if ((octet3 == '\0') || ((octet3 >> 6) != 2)) { *codepointSize = 4; return codepoint; }  // Unexpected sequence
 
         // [0]xF0       [1]x90-BF       [2]UTF8-tail  [3]UTF8-tail
         // [0]xF1-F3    [1]UTF8-tail    [2]UTF8-tail  [3]UTF8-tail
         // [0]xF4       [1]x80-8F       [2]UTF8-tail  [3]UTF8-tail
 
         if (((octet == 0xf0) && !((octet1 >= 0x90) && (octet1 <= 0xbf))) ||
-            ((octet == 0xf4) && !((octet1 >= 0x80) && (octet1 <= 0x8f)))) { *bytesProcessed = 2; return code; } // Unexpected sequence
+            ((octet == 0xf4) && !((octet1 >= 0x80) && (octet1 <= 0x8f)))) { *codepointSize = 2; return codepoint; } // Unexpected sequence
 
         if (octet >= 0xf0)
         {
-            code = ((octet & 0x7) << 18) | ((octet1 & 0x3f) << 12) | ((octet2 & 0x3f) << 6) | (octet3 & 0x3f);
-            *bytesProcessed = 4;
+            codepoint = ((octet & 0x7) << 18) | ((octet1 & 0x3f) << 12) | ((octet2 & 0x3f) << 6) | (octet3 & 0x3f);
+            *codepointSize = 4;
         }
     }
 
-    if (code > 0x10ffff) code = 0x3f;     // Codepoints after U+10ffff are invalid
+    if (codepoint > 0x10ffff) codepoint = 0x3f;     // Codepoints after U+10ffff are invalid
+
+    return codepoint;
+}
+
+// Get next codepoint in a byte sequence and bytes processed
+int GetCodepointNext(const char *text, int *codepointSize)
+{
+    const char *ptr = text;
+    int codepoint = 0x3f;       // Codepoint (defaults to '?')
+    *codepointSize = 0;
+    
+    // Get current codepoint and bytes processed
+    if (0xf0 == (0xf8 & ptr[0]))
+    {
+        // 4 byte UTF-8 codepoint
+        codepoint = ((0x07 & ptr[0]) << 18) | ((0x3f & ptr[1]) << 12) | ((0x3f & ptr[2]) << 6) | (0x3f & ptr[3]);
+        *codepointSize = 4;
+    }
+    else if (0xe0 == (0xf0 & ptr[0]))
+    {
+        // 3 byte UTF-8 codepoint */
+        codepoint = ((0x0f & ptr[0]) << 12) | ((0x3f & ptr[1]) << 6) | (0x3f & ptr[2]);
+        *codepointSize = 3;
+    } 
+    else if (0xc0 == (0xe0 & ptr[0]))
+    {
+        // 2 byte UTF-8 codepoint
+        codepoint = ((0x1f & ptr[0]) << 6) | (0x3f & ptr[1]);
+        *codepointSize = 2;
+    } 
+    else
+    {
+        // 1 byte UTF-8 codepoint
+        codepoint = ptr[0];
+        *codepointSize = 1;
+    }
+    
+    return codepoint;
+}
+
+// Get previous codepoint in a byte sequence and bytes processed
+int GetCodepointPrevious(const char *text, int *codepointSize)
+{
+    const char *ptr = text;
+    int codepoint = 0x3f;       // Codepoint (defaults to '?')
+    int cpSize = 0;
+    *codepointSize = 0;
+
+    // Move to previous codepoint
+    do ptr--;
+    while (((0x80 & ptr[0]) != 0) && ((0xc0 & ptr[0]) ==  0x80));
+    
+    codepoint = GetCodepointNext(ptr, &cpSize);
+
+    if (codepoint != 0) *codepointSize = cpSize;
 
-    return code;
+    return codepoint;
 }
 
 //----------------------------------------------------------------------------------
diff --git a/src/rtextures.c b/src/rtextures.c
index 3badc349..41bd3e18 100644
--- a/src/rtextures.c
+++ b/src/rtextures.c
@@ -1238,8 +1238,8 @@ Image ImageTextEx(Font font, const char *text, float fontSize, float spacing, Co
     {
         // Get next codepoint from byte string and glyph index in font
         int codepointByteCount = 0;
-        int codepoint = GetCodepoint(&text[i], &codepointByteCount);    // WARNING: Module required: rtext
-        int index = GetGlyphIndex(font, codepoint);                     // WARNING: Module required: rtext
+        int codepoint = GetCodepointNext(&text[i], &codepointByteCount);    // WARNING: Module required: rtext
+        int index = GetGlyphIndex(font, codepoint);                         // WARNING: Module required: rtext
 
         // NOTE: Normally we exit the decoding sequence as soon as a bad byte is found (and return 0x3f)
         // but we need to draw all of the bad bytes using the '?' symbol moving one byte