lib: unichar - Add uni_utf8_get_char_buf()

Unlike the existing alternatives, this will not stop reading input at '\0'.

diff --git a/src/lib/unichar.c b/src/lib/unichar.c
index 8efacffc09e6aad627029af93c23bd2dc23d15e0..70db1c73507372aae2b8cebd073c8f91f8226cf9 100644 (file)
--- a/src/lib/unichar.c
+++ b/src/lib/unichar.c
@@ -29,22 +29,18 @@ unsigned int uni_strlen(const unichar_t *str)
        return len;
 }
 
-int uni_utf8_get_char(const char *input, unichar_t *chr_r)
-{
-       return uni_utf8_get_char_n((const unsigned char *)input, SIZE_MAX,
-                                  chr_r);
-}
-
-int uni_utf8_get_char_n(const void *_input, size_t max_len, unichar_t *chr_r)
+static int
+uni_utf8_parse_char(const void *_buffer, size_t size, bool cstr,
+                   unichar_t *chr_r)
 {
        static unichar_t lowest_valid_chr_table[] =
                { 0, 0, 0x80, 0x800, 0x10000, 0x200000, 0x4000000 };
-       const unsigned char *input = _input;
+       const unsigned char *input = _buffer;
        unichar_t chr, lowest_valid_chr;
        unsigned int i, len;
        int ret;
 
-       i_assert(max_len > 0);
+       i_assert(size > 0);
 
        if (*input < 0x80) {
                *chr_r = *input;
@@ -77,20 +73,20 @@ int uni_utf8_get_char_n(const void *_input, size_t max_len, unichar_t *chr_r)
                return -1;
        }
 
-       if (len <= max_len) {
+       if (len <= size) {
                lowest_valid_chr = lowest_valid_chr_table[len];
                ret = len;
        } else {
                /* check first if the input is invalid before returning 0 */
                lowest_valid_chr = 0;
                ret = 0;
-               len = max_len;
+               len = size;
        }
 
        /* the following bytes must all be 10xxxxxx */
        for (i = 1; i < len; i++) {
                if ((input[i] & 0xc0) != 0x80) {
-                       return (max_len == SIZE_MAX && input[i] == '\0' ?
+                       return (cstr && size == SIZE_MAX && input[i] == '\0' ?
                                0 : -1);
                }
 
@@ -110,6 +106,21 @@ int uni_utf8_get_char_n(const void *_input, size_t max_len, unichar_t *chr_r)
        return ret;
 }
 
+int uni_utf8_get_char(const char *input, unichar_t *chr_r)
+{
+       return uni_utf8_parse_char(input, SIZE_MAX, TRUE, chr_r);
+}
+
+int uni_utf8_get_char_n(const void *input, size_t max_len, unichar_t *chr_r)
+{
+       return uni_utf8_parse_char(input, max_len, TRUE, chr_r);
+}
+
+int uni_utf8_get_char_buf(const void *buffer, size_t size, unichar_t *chr_r)
+{
+       return uni_utf8_parse_char(buffer, size, FALSE, chr_r);
+}
+
 int uni_utf8_to_ucs4(const char *input, ARRAY_TYPE(unichars) *output)
 {
        unichar_t chr;

diff --git a/src/lib/unichar.h b/src/lib/unichar.h
index d75a6ba34c3c045b26871cfdec02080ad410fe52..e26acaf298f86ca5e2d54c5fff0ac63e8539d9b7 100644 (file)
--- a/src/lib/unichar.h
+++ b/src/lib/unichar.h
@@ -89,6 +89,7 @@ static inline unsigned int uni_ucs4_to_utf8_len(unichar_t chr)
    -1 for invalid input. */
 int uni_utf8_get_char(const char *input, unichar_t *chr_r);
 int uni_utf8_get_char_n(const void *input, size_t max_len, unichar_t *chr_r);
+int uni_utf8_get_char_buf(const void *buffer, size_t size, unichar_t *chr_r);
 /* Returns number of characters in UTF-8 string. */
 unsigned int uni_utf8_strlen(const char *input) ATTR_PURE;
 /* Returns number of characters in UTF-8 input of specified size. */