Merge #8635 from janlazo/vim-8.0.0252

src/nvim/charset.c

@@ -826,7 +826,7 @@ bool vim_isIDc(int c)
 /// For multi-byte characters mb_get_class() is used (builtin rules).
 ///
 /// @param  c  character to check
-bool vim_iswordc(int c)
+bool vim_iswordc(const int c)
   FUNC_ATTR_PURE FUNC_ATTR_WARN_UNUSED_RESULT
 {
   return vim_iswordc_buf(c, curbuf);
@@ -842,7 +842,7 @@ bool vim_iswordc_tab(const int c, const uint64_t *const chartab)
   FUNC_ATTR_PURE FUNC_ATTR_WARN_UNUSED_RESULT FUNC_ATTR_NONNULL_ALL
 {
   return (c >= 0x100
-          ? (utf_class(c) >= 2)
+          ? (utf_class_tab(c, chartab) >= 2)
           : (c > 0 && GET_CHARTAB_TAB(chartab, c) != 0));
 }
 
@@ -852,7 +852,7 @@ bool vim_iswordc_tab(const int c, const uint64_t *const chartab)
 ///
 /// @param  c    character to check
 /// @param  buf  buffer whose keywords to use
-bool vim_iswordc_buf(int c, buf_T *buf)
+bool vim_iswordc_buf(const int c, buf_T *const buf)
   FUNC_ATTR_PURE FUNC_ATTR_WARN_UNUSED_RESULT FUNC_ATTR_NONNULL_ARG(2)
 {
   return vim_iswordc_tab(c, buf->b_chartab);
@@ -863,13 +863,10 @@ bool vim_iswordc_buf(int c, buf_T *buf)
 /// @param  p  pointer to the multi-byte character
 ///
 /// @return true if "p" points to a keyword character.
-bool vim_iswordp(char_u *p)
+bool vim_iswordp(const char_u *const p)
   FUNC_ATTR_PURE FUNC_ATTR_WARN_UNUSED_RESULT FUNC_ATTR_NONNULL_ALL
 {
-  if (MB_BYTE2LEN(*p) > 1) {
-    return mb_get_class(p) >= 2;
-  }
-  return GET_CHARTAB(curbuf, *p) != 0;
+  return vim_iswordp_buf(p, curbuf);
 }
 
 /// Just like vim_iswordc_buf() but uses a pointer to the (multi-byte)
@@ -879,13 +876,15 @@ bool vim_iswordp(char_u *p)
 /// @param  buf  buffer whose keywords to use
 ///
 /// @return true if "p" points to a keyword character.
-bool vim_iswordp_buf(char_u *p, buf_T *buf)
+bool vim_iswordp_buf(const char_u *const p, buf_T *const buf)
   FUNC_ATTR_PURE FUNC_ATTR_WARN_UNUSED_RESULT FUNC_ATTR_NONNULL_ALL
 {
-  if (MB_BYTE2LEN(*p) > 1) {
-    return mb_get_class(p) >= 2;
+  int c = *p;
+
+  if (MB_BYTE2LEN(c) > 1) {
+    c = utf_ptr2char(p);
   }
-  return GET_CHARTAB(buf, *p) != 0;
+  return vim_iswordc_buf(c, buf);
 }
 
 /// Check that "c" is a valid file-name character.

src/nvim/mbyte.c

@@ -428,7 +428,7 @@ int mb_get_class_tab(const char_u *p, const uint64_t *const chartab)
     }
     return 1;
   }
-  return utf_class(utf_ptr2char(p));
+  return utf_class_tab(utf_ptr2char(p), chartab);
 }
 
 /*
@@ -1039,7 +1039,12 @@ bool utf_printable(int c)
  * 1: punctuation
  * 2 or bigger: some class of word character.
  */
-int utf_class(int c)
+int utf_class(const int c)
+{
+  return utf_class_tab(c, curbuf->b_chartab);
+}
+
+int utf_class_tab(const int c, const uint64_t *const chartab)
 {
   /* sorted list of non-overlapping intervals */
   static struct clinterval {
@@ -1122,11 +1127,13 @@ int utf_class(int c)
 
   /* First quick check for Latin1 characters, use 'iskeyword'. */
   if (c < 0x100) {
-    if (c == ' ' || c == '\t' || c == NUL || c == 0xa0)
-      return 0;             /* blank */
-    if (vim_iswordc(c))
-      return 2;             /* word character */
-    return 1;               /* punctuation */
+    if (c == ' ' || c == '\t' || c == NUL || c == 0xa0) {
+      return 0;             // blank
+    }
+    if (vim_iswordc_tab(c, chartab)) {
+      return 2;             // word character
+    }
+    return 1;               // punctuation
   }
 
   /* binary search in table */

test/unit/mbyte_spec.lua

@@ -5,6 +5,7 @@ local ffi     = helpers.ffi
 local eq      = helpers.eq
 
 local mbyte = helpers.cimport("./src/nvim/mbyte.h")
+local charset = helpers.cimport('./src/nvim/charset.h')
 
 describe('mbyte', function()
 
@@ -42,6 +43,15 @@ describe('mbyte', function()
     -- Sequences with more than four bytes
   end)
 
+  itp('utf_char2bytes', function()
+    local char_p = ffi.typeof('char[?]')
+    for c = 0, 0xFFFF do
+      local p = char_p(4, 0)
+      mbyte.utf_char2bytes(c, p)
+      eq(c, mbyte.utf_ptr2char(p))
+      eq(charset.vim_iswordc(c), charset.vim_iswordp(p))
+    end
+  end)
 
   describe('utfc_ptr2char_len', function()