void test_unichar(void)
{
static const char overlong_utf8[] = "\xf8\x80\x95\x81\xa1";
static const char collate_in[] = "\xc3\xbc \xc2\xb3";
static const char collate_exp[] = "U\xcc\x88 3";
buffer_t *collate_out;
unichar_t chr, chr2;
string_t *str = t_str_new(16);
test_begin("unichars encode/decode");
for (chr = 0; chr <= 0x10ffff; chr++) {
/* The bottom 6 bits should be irrelevant to code coverage,
only test 000000, 111111, and something in between. */
if ((chr & 63) == 1)
chr += rand() % 62; /* After 0, somewhere between 1 and 62 */
else if ((chr & 63) > 0 && (chr & 63) < 63)
chr |= 63; /* After random, straight to 63 */
str_truncate(str, 0);
uni_ucs4_to_utf8_c(chr, str);
test_assert(uni_utf8_str_is_valid(str_c(str)));
test_assert(uni_utf8_get_char(str_c(str), &chr2) == (int)uni_utf8_char_bytes(*str_data(str)));
test_assert(chr2 == chr);
if ((chr & 0x63) == 0) {
unsigned int utf8len = uni_utf8_char_bytes(*str_c(str));
/* virtually truncate the byte string */
while (--utf8len > 0)
test_assert(uni_utf8_get_char_n(str_c(str), utf8len, &chr2) == 0);
utf8len = uni_utf8_char_bytes(*str_c(str));
/* actually truncate the byte stream */
while (--utf8len > 0) {
str_truncate(str, utf8len);
test_assert(!uni_utf8_str_is_valid(str_c(str)));
test_assert(uni_utf8_get_char(str_c(str), &chr2) == 0);
}
}
}
test_end();
test_begin("unichar collation");
collate_out = buffer_create_dynamic(default_pool, 32);
uni_utf8_to_decomposed_titlecase(collate_in, sizeof(collate_in),
collate_out);
test_assert(!strcmp(collate_out->data, collate_exp));
buffer_free(&collate_out);
test_assert(!uni_utf8_str_is_valid(overlong_utf8));
test_assert(uni_utf8_get_char(overlong_utf8, &chr2) < 0);
test_end();
test_unichar_uni_utf8_strlen();
test_unichar_uni_utf8_partial_strlen_n();
}
void str_sanitize_append(string_t *dest, const char *src, size_t max_len)
{
unsigned int len;
unichar_t chr;
size_t i;
int ret;
for (i = 0; i < max_len && src[i] != '\0'; ) {
len = uni_utf8_char_bytes(src[i]);
ret = uni_utf8_get_char(src+i, &chr);
if (ret <= 0) {
/* invalid UTF-8 */
str_append_c(dest, '?');
if (ret == 0) {
/* input ended too early */
return;
}
i++;
continue;
}
if ((unsigned char)src[i] < 32)
str_append_c(dest, '?');
else
str_append_c(dest, src[i]);
i += len;
}
if (src[i] != '\0') {
str_truncate(dest, str_len(dest) <= 3 ? 0 : str_len(dest)-3);
str_append(dest, "...");
}
}
static void
xml_encode_data(string_t *dest, const unsigned char *data, unsigned int len)
{
unichar_t chr;
unsigned int i;
for (i = 0; i < len; i++) {
switch (data[i]) {
case '&':
str_append(dest, "&");
break;
case '<':
str_append(dest, "<");
break;
case '>':
str_append(dest, ">");
break;
case '\t':
case '\n':
case '\r':
/* exceptions to the following control char check */
str_append_c(dest, data[i]);
break;
default:
if (data[i] < 32) {
/* SOLR doesn't like control characters.
replace them with spaces. */
str_append_c(dest, ' ');
} else if (data[i] >= 0x80) {
/* make sure the character is valid for XML
so we don't get XML parser errors */
unsigned int char_len =
uni_utf8_char_bytes(data[i]);
if (i + char_len <= len &&
uni_utf8_get_char_n(data + i, char_len, &chr) == 1 &&
is_valid_xml_char(chr))
str_append_n(dest, data + i, char_len);
else {
str_append_n(dest, utf8_replacement_char,
UTF8_REPLACEMENT_CHAR_LEN);
}
i += char_len - 1;
} else {
str_append_c(dest, data[i]);
}
break;
}
}
}
static size_t str_sanitize_skip_start(const char *src, size_t max_len)
{
unsigned int len;
unichar_t chr;
size_t i;
for (i = 0; i < max_len; ) {
len = uni_utf8_char_bytes(src[i]);
if (uni_utf8_get_char(src+i, &chr) <= 0)
break;
if ((unsigned char)src[i] < 32)
break;
i += len;
}
return i;
}
static size_t str_sanitize_skip_start(const char *src, size_t max_bytes)
{
unsigned int len;
unichar_t chr;
size_t i;
for (i = 0; src[i] != '\0'; ) {
len = uni_utf8_char_bytes(src[i]);
if (i + len > max_bytes || uni_utf8_get_char(src+i, &chr) <= 0)
break;
if ((unsigned char)src[i] < 32)
break;
i += len;
}
i_assert(i <= max_bytes);
return i;
}
void
fts_tokenizer_delete_trailing_partial_char(const unsigned char *data,
size_t *len)
{
size_t pos;
unsigned int char_bytes;
/* the token is truncated - make sure the last character
exists entirely in the token */
for (pos = *len-1; pos > 0; pos--) {
if (UTF8_IS_START_SEQ(data[pos]))
break;
}
char_bytes = uni_utf8_char_bytes(data[pos]);
if (char_bytes != *len-pos) {
i_assert(char_bytes > *len-pos);
*len = pos;
}
}
void str_sanitize_append(string_t *dest, const char *src, size_t max_bytes)
{
unsigned int len, initial_pos = str_len(dest);
unichar_t chr;
size_t i;
int ret;
for (i = 0; src[i] != '\0'; ) {
len = uni_utf8_char_bytes(src[i]);
if (i + len > max_bytes)
break;
ret = uni_utf8_get_char(src+i, &chr);
if (ret <= 0) {
/* invalid UTF-8 */
str_append_c(dest, '?');
if (ret == 0) {
/* input ended too early */
return;
}
i++;
continue;
}
if ((unsigned char)src[i] < 32)
str_append_c(dest, '?');
else
str_append_n(dest, src+i, len);
i += len;
}
if (src[i] != '\0') {
if (max_bytes < 3)
str_truncate(dest, initial_pos);
else {
while (str_len(dest) - initial_pos > max_bytes-3)
str_sanitize_truncate_char(dest, initial_pos);
}
str_append(dest, "...");
}
}
int uni_utf8_get_char_n(const void *_input, size_t max_len, unichar_t *chr_r)
{
static unichar_t lowest_valid_chr_table[] =
{ 0, 0, 0x80, 0x800, 0x10000, 0x200000, 0x4000000 };
const unsigned char *input = _input;
unichar_t chr, lowest_valid_chr;
unsigned int i, len;
int ret;
i_assert(max_len > 0);
if (*input < 0x80) {
*chr_r = *input;
return 1;
}
/* first byte has len highest bits set, followed by zero bit.
the rest of the bits are used as the highest bits of the value. */
chr = *input;
len = uni_utf8_char_bytes(*input);
switch (len) {
case 2:
chr &= 0x1f;
break;
case 3:
chr &= 0x0f;
break;
case 4:
chr &= 0x07;
break;
case 5:
chr &= 0x03;
break;
case 6:
chr &= 0x01;
break;
default:
/* only 7bit chars should have len==1 */
i_assert(len == 1);
return -1;
}
if (len <= max_len) {
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;
}
/* the following bytes must all be 10xxxxxx */
for (i = 1; i < len; i++) {
if ((input[i] & 0xc0) != 0x80)
return input[i] == '\0' ? 0 : -1;
chr <<= 6;
chr |= input[i] & 0x3f;
}
if (chr < lowest_valid_chr) {
/* overlong encoding */
return -1;
}
*chr_r = chr;
return ret;
}
static bool snippet_generate(struct snippet_context *ctx,
const unsigned char *data, size_t size)
{
size_t i, count;
if (ctx->html2text != NULL) {
buffer_set_used_size(ctx->plain_output, 0);
mail_html2text_more(ctx->html2text, data, size,
ctx->plain_output);
data = ctx->plain_output->data;
size = ctx->plain_output->used;
}
/* message-decoder should feed us only valid and complete
UTF-8 input */
for (i = 0; i < size; i += count) {
count = 1;
switch (ctx->state) {
case SNIPPET_STATE_NEWLINE:
if (data[i] == '>' && ctx->html2text == NULL) {
ctx->state = SNIPPET_STATE_QUOTED;
break;
}
ctx->state = SNIPPET_STATE_NORMAL;
/* fallthrough */
case SNIPPET_STATE_NORMAL:
if (size-i >= 3 &&
((data[i] == 0xEF && data[i+1] == 0xBB && data[i+2] == 0xBF) ||
(data[i] == 0xBF && data[i+1] == 0xBB && data[i+2] == 0xEF))) {
count += 2; /* because we skip +1 next */
break;
}
if (data[i] == '\r' || data[i] == '\n' ||
data[i] == '\t' || data[i] == ' ') {
/* skip any leading whitespace */
if (str_len(ctx->snippet) > 1)
ctx->add_whitespace = TRUE;
if (data[i] == '\n')
ctx->state = SNIPPET_STATE_NEWLINE;
break;
}
if (ctx->add_whitespace) {
str_append_c(ctx->snippet, ' ');
ctx->add_whitespace = FALSE;
if (ctx->chars_left-- == 0)
return FALSE;
}
if (ctx->chars_left-- == 0)
return FALSE;
count = uni_utf8_char_bytes(data[i]);
i_assert(i + count <= size);
str_append_n(ctx->snippet, data + i, count);
break;
case SNIPPET_STATE_QUOTED:
if (data[i] == '\n')
ctx->state = SNIPPET_STATE_NEWLINE;
break;
}
}
return TRUE;
}
