X-Git-Url: http://www.chiark.greenend.org.uk/ucgi/~mdw/git/disorder/blobdiff_plain/0b7052dac24e57edb180f07c0bd3479d397ebb74..caecd4f412422d8130d616da33325e03557218e7:/lib/unicode.h?ds=sidebyside

diff --git a/lib/unicode.h b/lib/unicode.h
index 08812b3..47b39f0 100644
--- a/lib/unicode.h
+++ b/lib/unicode.h
@@ -24,8 +24,21 @@
 #ifndef UNICODE_H
 #define UNICODE_H
 
+/** @brief Smart pointer into a string
+ *
+ * Iterators can be efficiently moved either forwards or back to the start of
+ * the string.  They cannot (currently) efficiently be moved backwards.  Their
+ * advantage is that they remember internal state to speed up boundary
+ * detection.
+ *
+ * Iterators can point to any code point of the string, or to a hypothetical
+ * post-final code point of value 0, but not outside the string.
+ */
+typedef struct utf32_iterator_data *utf32_iterator;
+
 char *utf32_to_utf8(const uint32_t *s, size_t ns, size_t *nd);
 uint32_t *utf8_to_utf32(const char *s, size_t ns, size_t *nd);
+int utf8_valid(const char *s, size_t ns);
 
 size_t utf32_len(const uint32_t *s);
 int utf32_cmp(const uint32_t *a, const uint32_t *b);
@@ -42,9 +55,39 @@ char *utf8_casefold_canon(const char *s, size_t ns, size_t *ndp);
 uint32_t *utf32_casefold_compat(const uint32_t *s, size_t ns, size_t *ndp);
 char *utf8_casefold_compat(const char *s, size_t ns, size_t *ndp);
 
-int utf32_is_gcb(const uint32_t *s, size_t ns, size_t n);
+int utf32_is_grapheme_boundary(const uint32_t *s, size_t ns, size_t n);
 int utf32_is_word_boundary(const uint32_t *s, size_t ns, size_t n);
 
+utf32_iterator utf32_iterator_new(const uint32_t *s, size_t ns);
+void utf32_iterator_destroy(utf32_iterator it);
+
+size_t utf32_iterator_where(utf32_iterator it);
+int utf32_iterator_set(utf32_iterator it, size_t n);
+int utf32_iterator_advance(utf32_iterator it, size_t n);
+uint32_t utf32_iterator_code(utf32_iterator it);
+int utf32_iterator_grapheme_boundary(utf32_iterator it);
+int utf32_iterator_word_boundary(utf32_iterator it);
+
+/** @brief Convert 0-terminated UTF-32 to UTF-8
+ * @param s 0-terminated UTF-32 string
+ * @return 0-terminated UTF-8 string or 0 on error
+ *
+ * See utf32_to_utf8() for possible causes of errors.
+ */
+static inline char *utf32nt_to_utf8(const uint32_t *s) {
+  return utf32_to_utf8(s, utf32_len(s), 0);
+}
+
+/** @brief Convert 0-terminated UTF-8 to UTF-32
+ * @param s 0-terminated UTF-8 string
+ * @return 0-terminated UTF-32 string or 0 on error
+ *
+ * See utf8_to_utf32() for possible causes of errors.
+ */
+static inline uint32_t *utf8nt_to_utf32(const char *s) {
+  return utf8_to_utf32(s, strlen(s), 0);
+}
+
 #endif /* UNICODE_H */
 
 /*