Changeset 28a5ebd in mainline for boot/generic/src/str.c

Timestamp:

2020-06-18T15:39:50Z (4 years ago)

Author:

Martin Decky <martin@…>

Branches:

lfn, master, serial, ticket/834-toolchain-update, topic/msim-upgrade, topic/simplify-dev-export

Children:

ce52c333

Parents:

4f663f3e

Message:

Use char32_t instead of wchat_t to represent UTF-32 strings

The intention of the native HelenOS string API has been always to
support Unicode in the UTF-8 and UTF-32 encodings as the sole character
representations and ignore the obsolete mess of older single-byte and
multibyte character encodings. Before C11, the wchar_t type has been
slightly misused for the purpose of the UTF-32 strings. The newer
char32_t type is obviously a much more suitable option. The standard
defines char32_t as uint_least32_t, thus we can take the liberty to fix
it to uint32_t.

To maintain compatilibity with the C Standard, the putwchar(wchar_t)
functions has been replaced by our custom putuchar(char32_t) functions
where appropriate.

File:

• boot/generic/src/str.c (modified) (13 diffs)

boot/generic/src/str.c

@@ -38 +38 @@
  * strings, called just strings are encoded in UTF-8. Wide strings (encoded
  * in UTF-32) are supported to a limited degree. A single character is
- * represented as wchar_t.@n
+ * represented as char32_t.@n
  *
  * Overview of the terminology:@n
@@ -46 +46 @@
  *  byte                  8 bits stored in uint8_t (unsigned 8 bit integer)
  *
- *  character             UTF-32 encoded Unicode character, stored in wchar_t
- *                        (signed 32 bit integer), code points 0 .. 1114111
+ *  character             UTF-32 encoded Unicode character, stored in char32_t
+ *                        (unsigned 32 bit integer), code points 0 .. 1114111
  *                        are valid
  *
@@ -57 +57 @@
  *
  *  wide string           UTF-32 encoded NULL-terminated Unicode string,
- *                        wchar_t *
+ *                        char32_t *
  *
  *  [wide] string size    number of BYTES in a [wide] string (excluding
@@ -96 +96 @@
  * A specific character inside a [wide] string can be referred to by:@n
  *
- *  pointer (char *, wchar_t *)
+ *  pointer (char *, char32_t *)
  *  byte offset (size_t)
  *  character index (size_t)
@@ -109 +109 @@
 #include <stdint.h>
 
-/** Check the condition if wchar_t is signed */
-#ifdef __WCHAR_UNSIGNED__
-#define WCHAR_SIGNED_CHECK(cond)  (true)
-#else
-#define WCHAR_SIGNED_CHECK(cond)  (cond)
-#endif
-
 /** Byte mask consisting of lowest @n bits (out of 8) */
 #define LO_MASK_8(n)  ((uint8_t) ((1 << (n)) - 1))
@@ -143 +136 @@
  *
  */
-wchar_t str_decode(const char *str, size_t *offset, size_t size)
+char32_t str_decode(const char *str, size_t *offset, size_t size)
 {
         if (*offset + 1 > size)
@@ -180 +173 @@
                 return U_SPECIAL;
 
-        wchar_t ch = b0 & LO_MASK_8(b0_bits);
+        char32_t ch = b0 & LO_MASK_8(b0_bits);
 
         /* Decode continuation bytes */
@@ -191 +184 @@
 
                 /* Shift data bits to ch */
-                ch = (ch << CONT_BITS) | (wchar_t) (b & LO_MASK_8(CONT_BITS));
+                ch = (ch << CONT_BITS) | (char32_t) (b & LO_MASK_8(CONT_BITS));
                 cbytes--;
         }
@@ -213 +206 @@
  *         code was invalid.
  */
-errno_t chr_encode(const wchar_t ch, char *str, size_t *offset, size_t size)
+errno_t chr_encode(const char32_t ch, char *str, size_t *offset, size_t size)
 {
         if (*offset >= size)
@@ -340 +333 @@
  *
  */
-bool ascii_check(wchar_t ch)
-{
-        if (WCHAR_SIGNED_CHECK(ch >= 0) && (ch <= 127))
+bool ascii_check(char32_t ch)
+{
+        if (ch <= 127)
                 return true;
 
@@ -353 +346 @@
  *
  */
-bool chr_check(wchar_t ch)
-{
-        if (WCHAR_SIGNED_CHECK(ch >= 0) && (ch <= 1114111))
+bool chr_check(char32_t ch)
+{
+        if (ch <= 1114111)
                 return true;
 
@@ -381 +374 @@
 int str_cmp(const char *s1, const char *s2)
 {
-        wchar_t c1 = 0;
-        wchar_t c2 = 0;
+        char32_t c1 = 0;
+        char32_t c2 = 0;
 
         size_t off1 = 0;
@@ -421 +414 @@
         size_t dest_off = 0;
 
-        wchar_t ch;
+        char32_t ch;
         while ((ch = str_decode(src, &src_off, STR_NO_LIMIT)) != 0) {
                 if (chr_encode(ch, dest, &dest_off, size - 1) != EOK)