FIX: length? of any-word!

fixes: Oldes/Rebol-issues#1740
fixes: Oldes/Rebol-issues#2224

src/core/s-unicode.c

@@ -972,6 +972,20 @@ ConversionResult ConvertUTF8toUTF32 (
  return dst;
 }
 
+/***********************************************************************
+**
+*/ REBCNT Length_As_UTF8_Code_Points(REBYTE *src)
+/*
+** Returns number of code points encoded in UTF-8.
+**
+***********************************************************************/
+{
+ REBCNT size = 0;
+ while (*src) {
+ size += (*src++ & 0xC0) != 0x80;
+ }
+ return size;
+}
 
 /***********************************************************************
 **

src/core/t-word.c

@@ -84,8 +84,8 @@
 
  switch (action) {
  case A_LENGTHQ:
- diff = (REBINT)LEN_BYTES(Get_Sym_Name(VAL_WORD_SYM(val)));
- if (type != REB_WORD) diff++;
+ diff = (REBINT)Length_As_UTF8_Code_Points(Get_Sym_Name(VAL_WORD_SYM(val)));
+ //if (type != REB_WORD) diff++; // in case that the _decoration_ should be also counted (#abc :abc abc: 'abc)
  DS_Ret_Int(diff);
  break;
 

src/tests/run-tests.r3

@@ -40,6 +40,7 @@ dt [ ;- delta time
  wrap load %units/tuple-test.r3
  wrap load %units/typeset-test.r3
  wrap load %units/vector-test.r3
+ wrap load %units/word-test.r3
  ;- cryptography tests:
  wrap load %units/aes-test.r3
  wrap load %units/chacha20-test.r3

src/tests/units/word-test.r3

@@ -0,0 +1,25 @@
+Rebol [
+ Title: "Rebol3 word test script"
+ Author: "Oldes, Peter W A Wood"
+ File: %word-test.r3
+ Tabs: 4
+ Needs: [%../quick-test-module.r3]
+]
+
+~~~start-file~~~ "WORD!"
+
+===start-group=== "word"
+ --test-- "Length? any-word!"
+ ;@@ https://github.com/Oldes/Rebol-issues/issues/1740
+ ;@@ https://github.com/Oldes/Rebol-issues/issues/2224
+ --assert 1 = length? #a
+ --assert 1 = length? 'a
+ --assert 1 = length? quote 'a
+ --assert 2 = length? #ša
+ --assert 2 = length? quote 'ša
+ --assert 1 = length? to-word to-string to-char 126
+ --assert 1 = length? to-word to-string to-char 128
+
+===end-group===
+
+~~~end-file~~~