Added unicode encoding example in C

1 files changed, 54 insertions, 0 deletions

diff --git a/all_unicode_chars/all_unicode_chars.c b/all_unicode_chars/all_unicode_chars.c
new file mode 100644
index 0000000..33099b1
--- /dev/null
+++ b/all_unicode_chars/all_unicode_chars.c
@@ -0,0 +1,54 @@
+#include <stdlib.h>
+#include <stdio.h>
+#include <stdint.h>
+#include <unistd.h>
+#include <errno.h>
+#include <locale.h>
+#include <string.h>
+#include <uchar.h>
+#include <wchar.h>
+
+int main( int argc, char *args[] ){
+    setlocale( LC_ALL, "" );
+    fprintf( stdout, "Buffer size: %llu\n", (0xFFFF - (0x00A0 - 0x007F)) * 4 );
+    uint16_t char_buffer[(0xFFFF - (0x00A0 - 0x007F)) * 4] = {0};
+    memset( (void*)&char_buffer, 0, (0xFFFF - (0x00A0 - 0x007F)) * 4 );
+    uint16_t current_char = 0x0020;
+    size_t char_bytes = 0;
+    char mbstr[sizeof(wchar_t)] = {0};
+    memset( (void*)&mbstr, 0, sizeof(wchar_t) );
+    size_t b_size = 0;
+    mbstate_t *m_state = NULL;
+    for ( ; current_char < 0xFFFF; current_char++ ){
+        if ( current_char > 0x007F && current_char < 0x00A0 )
+            current_char = 0x00A0;
+        if ( (b_size = wcrtomb((char*)&mbstr, (wchar_t)current_char, m_state)) == (size_t)-1 )
+            continue;
+        wmemmove( (void*)&char_buffer[char_bytes], (const void*)&mbstr, b_size );
+        memset( (void*)&mbstr, 0, b_size );
+        char_bytes += b_size;
+    }
+    fprintf( stdout, "Copied %llu bytes into buffer\n", char_bytes );
+    b_size = 0;
+    m_state = NULL;
+    memset( (void*)&mbstr, 0, sizeof(mbstr) );
+    FILE *output = fopen( "unicode_chars.txt", "w" );
+    if ( output == NULL ){
+        fprintf( stderr, "Error: %s\n", strerror(errno) );
+        exit( EXIT_FAILURE );
+    }
+    ssize_t char_written = 0;
+    if ( (char_written = (ssize_t)fwrite((void*)&char_buffer, sizeof(uint16_t), char_bytes, output)) < 0 ){
+        fclose( output );
+        fprintf( stderr, "Error writing file: %s\n", strerror(errno) );
+        exit( EXIT_FAILURE );
+    }
+    fclose( output );
+    fprintf( stdout, "%llu bytes written into unicode_chars.txt\n", char_written );
+    memset( (void*)&char_buffer, 0, sizeof(char_buffer) );
+    output = NULL;
+    current_char = 0x0020;
+    char_written = 0;
+    char_bytes = 0;
+    exit( EXIT_SUCCESS );
+}