andreacavalli/tdlib-fork
Archived
1
0
Fork 0
Code Pull Requests Releases Activity

Update strip_empty_characters.

Browse Source 
GitOrigin-RevId: 480beb7b352b28f59f65a63fd1d4550d8e16803f
This commit is contained in:
levlam 2020-05-07 01:21:05 +03:00
parent c376c1ac08
commit 4d9b8cf016
2 changed files with 19 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

20
td/telegram/misc.cpp
View File

@ -156,8 +156,8 @@ bool clean_input_string(string &str) {
string strip_empty_characters(string str, size_t max_length, bool strip_rtlo) { string strip_empty_characters(string str, size_t max_length, bool strip_rtlo) {
static const char *space_characters[] = {u8"\u1680", u8"\u180E", u8"\u2000", u8"\u2001", u8"\u2002", static const char *space_characters[] = {u8"\u1680", u8"\u180E", u8"\u2000", u8"\u2001", u8"\u2002",
u8"\u2003", u8"\u2004", u8"\u2005", u8"\u2006", u8"\u2007", u8"\u2003", u8"\u2004", u8"\u2005", u8"\u2006", u8"\u2007",
u8"\u2008", u8"\u2009", u8"\u200A", u8"\u200B", u8"\u202E", u8"\u2008", u8"\u2009", u8"\u200A", u8"\u202E", u8"\u202F",
u8"\u202F", u8"\u205F", u8"\u3000", u8"\uFEFF", u8"\uFFFC"}; u8"\u205F", u8"\u2800", u8"\u3000", u8"\uFFFC"};
static bool can_be_first[std::numeric_limits<unsigned char>::max() + 1]; static bool can_be_first[std::numeric_limits<unsigned char>::max() + 1];
static bool can_be_first_inited = [&] { static bool can_be_first_inited = [&] {
for (auto space_ch : space_characters) { for (auto space_ch : space_characters) {
@ -197,9 +197,13 @@ string strip_empty_characters(string str, size_t max_length, bool strip_rtlo) {
Slice trimmed = trim(utf8_truncate(trim(Slice(str.c_str(), new_len)), max_length)); Slice trimmed = trim(utf8_truncate(trim(Slice(str.c_str(), new_len)), max_length));
// check if there is some non-empty character, empty characters: // check if there is some non-empty character, empty characters:
// "\xE2\x80\x8B", ZERO WIDTH SPACE
// "\xE2\x80\x8C", ZERO WIDTH NON-JOINER // "\xE2\x80\x8C", ZERO WIDTH NON-JOINER
// "\xE2\x80\x8D", ZERO WIDTH JOINER // "\xE2\x80\x8D", ZERO WIDTH JOINER
// "\xE2\x80\x8E", LEFT-TO-RIGHT MARK
// "\xE2\x80\x8F", RIGHT-TO-LEFT MARK
// "\xE2\x80\xAE", RIGHT-TO-LEFT OVERRIDE // "\xE2\x80\xAE", RIGHT-TO-LEFT OVERRIDE
// "\xEF\xBB\xBF", ZERO WIDTH NO-BREAK SPACE aka BYTE ORDER MARK
// "\xC2\xA0", NO-BREAK SPACE // "\xC2\xA0", NO-BREAK SPACE
for (i = 0;;) { for (i = 0;;) {
if (i == trimmed.size()) { if (i == trimmed.size()) {
@ -211,9 +215,15 @@ string strip_empty_characters(string str, size_t max_length, bool strip_rtlo) {
i++; i++;
continue; continue;
} }
if (static_cast<unsigned char>(trimmed[i]) == 0xE2 && static_cast<unsigned char>(trimmed[i + 1]) == 0x80 && if (static_cast<unsigned char>(trimmed[i]) == 0xE2 && static_cast<unsigned char>(trimmed[i + 1]) == 0x80) {
(static_cast<unsigned char>(trimmed[i + 2]) == 0x8C || static_cast<unsigned char>(trimmed[i + 2]) == 0x8D || auto next = static_cast<unsigned char>(trimmed[i + 2]);
static_cast<unsigned char>(trimmed[i + 2]) == 0xAE)) { if ((0x8B <= next && next <= 0x8F) || next == 0xAE) {
i += 3;
continue;
}
}
if (static_cast<unsigned char>(trimmed[i]) == 0xEF && static_cast<unsigned char>(trimmed[i + 1]) == 0xBB &&
static_cast<unsigned char>(trimmed[i + 2]) == 0xBF) {
i += 3; i += 3;
continue; continue;
} }

6
test/string_cleaning.cpp
View File

@ -85,11 +85,11 @@ TEST(StringCleaning, strip_empty_characters) {
check_strip_empty_characters("/abc", 0, ""); check_strip_empty_characters("/abc", 0, "");
check_strip_empty_characters("/abc", 10000000, "/abc"); check_strip_empty_characters("/abc", 10000000, "/abc");
string spaces = string spaces =
u8"\u1680\u180E\u2000\u2001\u2002\u2003\u2004\u2005\u2006\u2007\u2008\u2009\u200A\u200B\u202F\u205F\u3000\uFEFF" u8"\u1680\u180E\u2000\u2001\u2002\u2003\u2004\u2005\u2006\u2007\u2008\u2009\u200A\u202F\u205F\u2800\u3000\uFFFC"
u8"\uFFFC\uFFFC"; u8"\uFFFC";
string spaces_replace = " "; string spaces_replace = " ";
string rtlo = u8"\u202E"; string rtlo = u8"\u202E";
string empty = "\xE2\x80\x8C\xE2\x80\x8D\xE2\x80\xAE\xC2\xA0\xC2\xA0"; string empty = "\xE2\x80\x8B\xE2\x80\x8C\xE2\x80\x8D\xE2\x80\x8E\xE2\x80\x8F\xE2\x80\xAE\xC2\xA0\xC2\xA0";
check_strip_empty_characters(spaces, 1000000, ""); check_strip_empty_characters(spaces, 1000000, "");
check_strip_empty_characters(spaces + rtlo, 1000000, ""); check_strip_empty_characters(spaces + rtlo, 1000000, "");