Changeset 28a5ebd in mainline for uspace/app/bdsh

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functions has been replaced by our custom putuchar(char32_t) functions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.

Location:

uspace/app/bdsh

Files:

• cmds/modules/cat/cat.c (modified) (3 diffs)
• cmds/modules/mkdir/mkdir.c (modified) (1 diff)
• cmds/modules/printf/printf.c (modified) (4 diffs)
• compl.c (modified) (2 diffs)
• tok.c (modified) (6 diffs)

uspace/app/bdsh/cmds/modules/cat/cat.c

@@ -154 +154 @@
 }
 
-static void paged_char(wchar_t c)
+static void paged_char(char32_t c)
 {
         if (last_char_was_newline && number) {
@@ -160 +160 @@
                 printf("%6u  ", lineno);
         }
-        putwchar(c);
+        putuchar(c);
         last_char_was_newline = c == '\n';
         if (paging_enabled) {
@@ -269 +269 @@
                                         paged_char(((count + i + 1) & 0xf) == 0 ? '\n' : ' ');
                                 } else {
-                                        wchar_t c = str_decode(buff, &offset, bytes);
+                                        char32_t c = str_decode(buff, &offset, bytes);
                                         if (c == 0) {
                                                 /* Reached end of string */

uspace/app/bdsh/cmds/modules/mkdir/mkdir.c

@@ -108 +108 @@
                 while (true) {
                         size_t prev_off = off;
-                        wchar_t cur_char = str_decode(path, &off, STR_NO_LIMIT);
+                        char32_t cur_char = str_decode(path, &off, STR_NO_LIMIT);
                         if ((cur_char == 0) || (cur_char == U_SPECIAL)) {
                                 break;

uspace/app/bdsh/cmds/modules/printf/printf.c

@@ -68 +68 @@
  * @param arg string with data to print.
  */
-static int print_arg(wchar_t ch, const char *arg)
+static int print_arg(char32_t ch, const char *arg)
 {
         switch (ch) {
@@ -93 +93 @@
  * @param ch  Control character.
  */
-static int process_ctl(wchar_t ch)
+static int process_ctl(char32_t ch)
 {
         switch (ch) {
@@ -120 +120 @@
         char *fmt;
         size_t pos, fmt_sz;
-        wchar_t ch;
+        char32_t ch;
         bool esc_flag = false;
         unsigned int carg;     // Current argument
@@ -170 +170 @@
                                 break;
                         }
-                        putwchar(ch);
+                        putuchar(ch);
                         break;
 
                 emit:
-                        putwchar(ch);
+                        putuchar(ch);
                         esc_flag = false;
                 }

uspace/app/bdsh/compl.c

@@ -44 +44 @@
 #include "util.h"
 
-static errno_t compl_init(wchar_t *text, size_t pos, size_t *cstart, void **state);
+static errno_t compl_init(char32_t *text, size_t pos, size_t *cstart, void **state);
 static errno_t compl_get_next(void *state, char **compl);
 static void compl_fini(void *state);
@@ -94 +94 @@
  * Set up iterators in completion object, based on current token.
  */
-static errno_t compl_init(wchar_t *text, size_t pos, size_t *cstart, void **state)
+static errno_t compl_init(char32_t *text, size_t pos, size_t *cstart, void **state)
 {
         compl_t *cs = NULL;

uspace/app/bdsh/tok.c

@@ -36 +36 @@
 
 /* Forward declarations of static functions */
-static wchar_t tok_get_char(tokenizer_t *);
-static wchar_t tok_look_char(tokenizer_t *);
-static errno_t tok_push_char(tokenizer_t *, wchar_t);
+static char32_t tok_get_char(tokenizer_t *);
+static char32_t tok_look_char(tokenizer_t *);
+static errno_t tok_push_char(tokenizer_t *, char32_t);
 static errno_t tok_push_token(tokenizer_t *);
 static bool tok_pending_chars(tokenizer_t *);
@@ -92 +92 @@
 {
         errno_t rc;
-        wchar_t next_char;
+        char32_t next_char;
 
         /* Read the input line char by char and append tokens */
@@ -182 +182 @@
 {
         errno_t rc;
-        wchar_t next_char;
+        char32_t next_char;
 
         while ((next_char = tok_look_char(tok)) != 0) {
@@ -214 +214 @@
 
 /** Get a char from input, advancing the input position */
-wchar_t tok_get_char(tokenizer_t *tok)
+char32_t tok_get_char(tokenizer_t *tok)
 {
         tok->in_char_offset++;
@@ -221 +221 @@
 
 /** Get a char from input, while staying on the same input position */
-wchar_t tok_look_char(tokenizer_t *tok)
+char32_t tok_look_char(tokenizer_t *tok)
 {
         size_t old_offset = tok->in_offset;
         size_t old_char_offset = tok->in_char_offset;
-        wchar_t ret = tok_get_char(tok);
+        char32_t ret = tok_get_char(tok);
         tok->in_offset = old_offset;
         tok->in_char_offset = old_char_offset;
@@ -232 +232 @@
 
 /** Append a char to the end of the current token */
-errno_t tok_push_char(tokenizer_t *tok, wchar_t ch)
+errno_t tok_push_char(tokenizer_t *tok, char32_t ch)
 {
         return chr_encode(ch, tok->outbuf, &tok->outbuf_offset, tok->outbuf_size);