filterANSIEscapes(): Handle UTF-8 characters

This commit is contained in:
Eelco Dolstra 2020-11-16 16:26:29 +01:00
parent 0d6419ad87
commit ef84c780bb
No known key found for this signature in database
GPG key ID: 8170B4726D7198DE
2 changed files with 32 additions and 1 deletions

View file

@ -587,4 +587,14 @@ namespace nix {
ASSERT_EQ(filterANSIEscapes(s, true), "foo bar baz" );
}
TEST(filterANSIEscapes, utf8) {
ASSERT_EQ(filterANSIEscapes("foobar", true, 5), "fooba");
ASSERT_EQ(filterANSIEscapes("fóóbär", true, 6), "fóóbär");
ASSERT_EQ(filterANSIEscapes("fóóbär", true, 5), "fóóbä");
ASSERT_EQ(filterANSIEscapes("fóóbär", true, 3), "fóó");
ASSERT_EQ(filterANSIEscapes("f€€bär", true, 4), "f€€b");
ASSERT_EQ(filterANSIEscapes("f𐍈𐍈bär", true, 4), "f𐍈𐍈b");
}
}

View file

@ -1409,7 +1409,28 @@ std::string filterANSIEscapes(const std::string & s, bool filterAll, unsigned in
i++;
else {
t += *i++; w++;
w++;
// Copy one UTF-8 character.
if ((*i & 0xe0) == 0xc0) {
t += *i++;
if (i != s.end() && ((*i & 0xc0) == 0x80)) t += *i++;
} else if ((*i & 0xf0) == 0xe0) {
t += *i++;
if (i != s.end() && ((*i & 0xc0) == 0x80)) {
t += *i++;
if (i != s.end() && ((*i & 0xc0) == 0x80)) t += *i++;
}
} else if ((*i & 0xf8) == 0xf0) {
t += *i++;
if (i != s.end() && ((*i & 0xc0) == 0x80)) {
t += *i++;
if (i != s.end() && ((*i & 0xc0) == 0x80)) {
t += *i++;
if (i != s.end() && ((*i & 0xc0) == 0x80)) t += *i++;
}
}
} else
t += *i++;
}
}