bullet prooffed the utf8 translation

4coder_lib/4coder_utf8.h

@@ -40,6 +40,20 @@ typedef int32_t b32_4tech;
 #endif
 // standard preamble end 
 
+static u32_4tech cp_min_by_utf8_length[] = {
+    0x0,
+    0x0,
+    0x80,
+    0x800,
+    0x10000,
+};
+
+static u32_4tech surrogate_min = 0xD800;
+static u32_4tech surrogate_max = 0xDFFF;
+
+static u32_4tech nonchar_min = 0xFDD0;
+static u32_4tech nonchar_max = 0xFDEF;
+
 static b32_4tech
 codepoint_is_whitespace(u32_4tech codepoint){
     b32_4tech result = false;
@@ -76,6 +90,11 @@ utf8_to_u32_length_unchecked(u8_4tech *buffer, u32_4tech *length_out){
         *length_out = 4;
     }
     
+    if (result < cp_min_by_utf8_length[*length_out] || (result >= surrogate_min && result <= surrogate_max) || (result >= nonchar_min && result <= nonchar_max) || ((result & 0xFFFF) >= 0xFE)){
+        result = 0;
+        *length_out = 0;
+    }
+    
     return(result);
 }
 
@@ -104,9 +123,12 @@ utf8_to_u32(u8_4tech **buffer_ptr, u8_4tech *end){
     else if (buffer[0] < 0xF0){
         length = 3;
     }
-    else{
+    else if (buffer[0] < 0xF8){
         length = 4;
     }
+    else{
+        length = 0;
+    }
     
     for (u32_4tech i = 1; i < length; ++i){
         if ((buffer[i] & 0xC0) != 0x80){
@@ -145,6 +167,11 @@ utf8_to_u32(u8_4tech **buffer_ptr, u8_4tech *end){
             }break;
         }
         
+        if (result < cp_min_by_utf8_length[length] || (result >= surrogate_min && result <= surrogate_max) || (result >= nonchar_min && result <= nonchar_max) || ((result & 0xFFFF) >= 0xFE)){
+            result = 0;
+            length = 0;
+        }
+        
         *buffer_ptr = buffer + length;
     }
     else{
@@ -171,7 +198,7 @@ u32_to_utf8_unchecked(u32_4tech codepoint, u8_4tech *buffer, u32_4tech *length_o
         buffer[2] = (u8_4tech)(0x80 | (codepoint & 0x3F));
         *length_out = 3;
     }
-    else{
+    else if (codepoint <= 0x10FFFF){
         codepoint &= 0x001FFFFF;
         buffer[0] = (u8_4tech)(0xF0 | (codepoint >> 18));
         buffer[1] = (u8_4tech)(0x80 | ((codepoint >> 12) & 0x3F));
@@ -179,6 +206,9 @@ u32_to_utf8_unchecked(u32_4tech codepoint, u8_4tech *buffer, u32_4tech *length_o
         buffer[3] = (u8_4tech)(0x80 | (codepoint & 0x3F));
         *length_out = 4;
     }
+    else{
+        *length_out = 0;
+    }
 }
 
 static umem_4tech
@@ -223,7 +253,7 @@ utf8_to_utf16_minimal_checking(u16_4tech *dst, umem_4tech max_wchars, u8_4tech *
             codepoint |= ((u32_4tech)((s[2])&0x3F));
             utf8_size = 3;
         }
-        else{
+        else if (s[0] < 0xF8){
             if (limit > 3){
                 *error = true;
                 break;
@@ -235,6 +265,15 @@ utf8_to_utf16_minimal_checking(u16_4tech *dst, umem_4tech max_wchars, u8_4tech *
             codepoint |= ((u32_4tech)((s[3])&0x3F));
             utf8_size = 4;
         }
+        else{
+            *error = true;
+            break;
+        }
+        
+        if (codepoint < cp_min_by_utf8_length[utf8_size]){
+            *error = true;
+            break;
+        }
         
         s += utf8_size;
         limit -= utf8_size;

4ed_translation.cpp

@@ -44,16 +44,17 @@ struct Translation_Emits{
     u32 step_count;
 };
 
-#define SINGLE_BYTE_ERROR_CLASS max_u8
+#define ERROR_BYTE (max_u8-1)
+#define CONTINUATION_BYTE max_u8
 
 internal void
 translating_consume_byte(Translation_State *tran, u8 ch, u32 i, u32 size, Translation_Byte_Description *desc_out){
     desc_out->byte_class = 0;
-    if ((ch >= ' ' && ch < 0x7F) || ch == '\t' || ch == '\n' || ch == '\r'){
+    if (ch < 0x80){
         desc_out->byte_class = 1;
     }
     else if (ch < 0xC0){
-        desc_out->byte_class = SINGLE_BYTE_ERROR_CLASS;
+        desc_out->byte_class = CONTINUATION_BYTE;
     }
     else if (ch < 0xE0){
         desc_out->byte_class = 2;
@@ -61,9 +62,12 @@ translating_consume_byte(Translation_State *tran, u8 ch, u32 i, u32 size, Transl
     else if (ch < 0xF0){
         desc_out->byte_class = 3;
     }
-    else{
+    else if (ch < 0xF8){
         desc_out->byte_class = 4;
     }
+    else{
+        desc_out->byte_class = ERROR_BYTE;
+    }
     
     desc_out->prelim_emit_type = BufferModelUnit_None;
     desc_out->last_byte_handler = TranLBH_None;
@@ -75,7 +79,7 @@ translating_consume_byte(Translation_State *tran, u8 ch, u32 i, u32 size, Transl
         if (desc_out->byte_class == 1){
             desc_out->prelim_emit_type = BufferModelUnit_Codepoint;
         }
-        else if (desc_out->byte_class == 0 || desc_out->byte_class == SINGLE_BYTE_ERROR_CLASS){
+        else if (desc_out->byte_class == 0 || desc_out->byte_class == CONTINUATION_BYTE || desc_out->byte_class == ERROR_BYTE){
             desc_out->prelim_emit_type = BufferModelUnit_Numbers;
         }
         else{
@@ -83,7 +87,7 @@ translating_consume_byte(Translation_State *tran, u8 ch, u32 i, u32 size, Transl
         }
     }
     else{
-        if (desc_out->byte_class == SINGLE_BYTE_ERROR_CLASS){
+        if (desc_out->byte_class == CONTINUATION_BYTE){
             tran->fill_buffer[tran->fill_i] = ch;
             ++tran->fill_i;
             
@@ -138,8 +142,13 @@ translating_select_emit_rule_with_font(System_Functions *system, Render_Font *fo
     type_out->codepoint_length = 0;
     if (desc.prelim_emit_type == BufferModelUnit_Codepoint){
         u32 cp = utf8_to_u32_length_unchecked(tran->fill_buffer, &type_out->codepoint_length);
-        type_out->codepoint = cp;
-        if (!font_can_render(system, font, cp)){
+        if (type_out->codepoint_length != 0){
+            type_out->codepoint = cp;
+            if (!font_can_render(system, font, cp)){
+                type_out->emit_type = BufferModelUnit_Numbers;
+            }
+        }
+        else{
             type_out->emit_type = BufferModelUnit_Numbers;
         }
     }