Changeset 82374b2 in mainline

Timestamp:

2011-06-28T07:32:48Z (13 years ago)

Author:

Oleg Romanenko <romanenko.oleg@…>

Branches:

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

Children:

fcc3cd8

Parents:

298a6ce

Message:

Add utf16_to_str function doing conversion between UTF16 string and UTF8
with surrogate pairs support

Location:

uspace/lib/c

Files:

• generic/str.c (modified) (1 diff)
• include/str.h (modified) (1 diff)

uspace/lib/c/generic/str.c

@@ -587 +587 @@
         }
 
+        dest[dest_off] = '\0';
+        return rc;
+}
+
+/** Convert UTF16 string to string.
+ *
+ * Convert utf16 string @a src to string. The output is written to the buffer
+ * specified by @a dest and @a size. @a size must be non-zero and the string
+ * written will always be well-formed. Surrogate pairs also supported.
+ *
+ * @param dest  Destination buffer.
+ * @param size  Size of the destination buffer.
+ * @param src   Source utf16 string.
+ *
+ * @return EOK, if success, negative otherwise.
+ */
+int utf16_to_str(char *dest, size_t size, const uint16_t *src)
+{
+        size_t idx=0, dest_off=0;
+        wchar_t ch;
+        int rc = EOK;
+
+        /* There must be space for a null terminator in the buffer. */
+        assert(size > 0);
+
+        while (src[idx]) {
+                if ((src[idx] & 0xfc00) == 0xd800) {
+                        if (src[idx+1] && (src[idx+1] & 0xfc00) == 0xdc00) {
+                                ch = 0x10000;
+                                ch += (src[idx] & 0x03FF) << 10;
+                                ch += (src[idx+1] & 0x03FF);
+                                idx += 2;
+                        }
+                        else
+                                break;
+                } else {
+                        ch = src[idx];
+                        idx++;
+                }
+                rc = chr_encode(ch, dest, &dest_off, size-1);
+                if (rc != EOK)
+                        break;
+        }
         dest[dest_off] = '\0';
         return rc;

uspace/lib/c/include/str.h

@@ -78 +78 @@
 extern char *wstr_to_astr(const wchar_t *src);
 extern int str_to_wstr(wchar_t *dest, size_t dlen, const char *src);
+extern int utf16_to_str(char *dest, size_t size, const uint16_t *src);
 
 extern char *str_chr(const char *str, wchar_t ch);