Skip to content

DX-116527 Update CHR to work with unicode #141

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
lriggs wants to merge 2 commits into from
Closed

DX-116527 Update CHR to work with unicode #141

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
04858f8
DX-116527 Update CHR to work with unicode
lriggs Jun 2, 2026
445bf8a
Update error message for readability.
lriggs Jun 2, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
62 changes: 43 additions & 19 deletions cpp/src/gandiva/precompiled/string_ops.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1387,40 +1387,64 @@ gdv_int32 ascii_utf8(const char* data, gdv_int32 data_len) {
return static_cast<gdv_int32>(static_cast<signed char>(data[0]));
}

// Returns the ASCII character having the binary equivalent to A.
// If A is larger than 256 the result is equivalent to chr(A % 256).
// Returns the UTF-8 encoding of the Unicode code point A.
// Raises an error if A is not a valid code point, i.e. it is negative, greater
// than 0x10FFFF, or falls within the UTF-16 surrogate range 0xD800-0xDFFF.
FORCE_INLINE
const char* chr_int32(gdv_int64 context, gdv_int32 in, gdv_int32* out_len) {
in = in % 256;
*out_len = 1;

char* ret = reinterpret_cast<char*>(gdv_fn_context_arena_malloc(context, *out_len));
if (ret == nullptr) {
gdv_fn_context_set_error_msg(context, "Could not allocate memory for output string");
const char* chr_int64(gdv_int64 context, gdv_int64 in, gdv_int32* out_len) {
if (in < 0 || in > 0x10FFFF || (in >= 0xD800 && in <= 0xDFFF)) {
gdv_fn_context_set_error_msg(
context, "Input is not a valid Unicode code point.");
*out_len = 0;
return "";
}
ret[0] = char(in);
return ret;
}

// Returns the ASCII character having the binary equivalent to A.
// If A is larger than 256 the result is equivalent to chr(A % 256).
FORCE_INLINE
const char* chr_int64(gdv_int64 context, gdv_int64 in, gdv_int32* out_len) {
in = in % 256;
*out_len = 1;
if (in <= 0x7F) {
*out_len = 1;
} else if (in <= 0x7FF) {
*out_len = 2;
} else if (in <= 0xFFFF) {
*out_len = 3;
} else {
*out_len = 4;
}

char* ret = reinterpret_cast<char*>(gdv_fn_context_arena_malloc(context, *out_len));
if (ret == nullptr) {
gdv_fn_context_set_error_msg(context, "Could not allocate memory for output string");
*out_len = 0;
return "";
}
ret[0] = char(in);

switch (*out_len) {
case 1:
ret[0] = static_cast<char>(in);
break;
case 2:
ret[0] = static_cast<char>(0xC0 | (in >> 6));
ret[1] = static_cast<char>(0x80 | (in & 0x3F));
break;
case 3:
ret[0] = static_cast<char>(0xE0 | (in >> 12));
ret[1] = static_cast<char>(0x80 | ((in >> 6) & 0x3F));
ret[2] = static_cast<char>(0x80 | (in & 0x3F));
break;
case 4:
ret[0] = static_cast<char>(0xF0 | (in >> 18));
ret[1] = static_cast<char>(0x80 | ((in >> 12) & 0x3F));
ret[2] = static_cast<char>(0x80 | ((in >> 6) & 0x3F));
ret[3] = static_cast<char>(0x80 | (in & 0x3F));
break;
}
return ret;
}

// Returns the UTF-8 encoding of the Unicode code point A. See chr_int64.
FORCE_INLINE
const char* chr_int32(gdv_int64 context, gdv_int32 in, gdv_int32* out_len) {
return chr_int64(context, in, out_len);
}

FORCE_INLINE
const char* convert_fromUTF8_binary(gdv_int64 context, const char* bin_in, gdv_int32 len,
gdv_int32* out_len) {
Expand Down
109 changes: 68 additions & 41 deletions cpp/src/gandiva/precompiled/string_ops_test.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -56,11 +56,12 @@ TEST(TestStringOps, TestAscii) {
}

TEST(TestStringOps, TestChrBigInt) {
// CHR
// CHR returns the UTF-8 encoding of the given Unicode code point.
gandiva::ExecutionContext ctx;
uint64_t ctx_ptr = reinterpret_cast<gdv_int64>(&ctx);
int32_t out_len = 0;

// 1-byte ASCII code points.
auto out = chr_int32(ctx_ptr, 88, &out_len);
EXPECT_EQ(std::string(out, out_len), "X");

Expand All @@ -70,62 +71,88 @@ TEST(TestStringOps, TestChrBigInt) {
out = chr_int32(ctx_ptr, 49, &out_len);
EXPECT_EQ(std::string(out, out_len), "1");

out = chr_int64(ctx_ptr, 84, &out_len);
EXPECT_EQ(std::string(out, out_len), "T");
out = chr_int32(ctx_ptr, 33, &out_len);
EXPECT_EQ(std::string(out, out_len), "!");

out = chr_int32(ctx_ptr, 340, &out_len);
EXPECT_EQ(std::string(out, out_len), "T");
out = chr_int64(ctx_ptr, 0, &out_len);
EXPECT_EQ(std::string(out, out_len), std::string("\0", 1));

out = chr_int64(ctx_ptr, 256, &out_len);
EXPECT_EQ(std::strcmp(out, "\0"), 0);
// BACKSPACE
out = chr_int64(ctx_ptr, 8, &out_len);
EXPECT_EQ(std::string(out, out_len), "\b");

out = chr_int32(ctx_ptr, 33, &out_len);
EXPECT_EQ(std::string(out, out_len), "!");
// ESCAPE (ESC)
out = chr_int64(ctx_ptr, 27, &out_len);
EXPECT_EQ(std::string(out, out_len), "\x1B");

out = chr_int64(ctx_ptr, 46, &out_len);
EXPECT_EQ(std::string(out, out_len), ".");
// Highest 1-byte code point (U+007F, DELETE).
out = chr_int32(ctx_ptr, 0x7F, &out_len);
EXPECT_EQ(std::string(out, out_len), "\x7F");

out = chr_int32(ctx_ptr, 63, &out_len);
EXPECT_EQ(std::string(out, out_len), "?");
// 2-byte code points.
// Lowest 2-byte code point (U+0080).
out = chr_int32(ctx_ptr, 0x80, &out_len);
EXPECT_EQ(std::string(out, out_len), "\xC2\x80");

out = chr_int64(ctx_ptr, 0, &out_len);
EXPECT_EQ(std::strcmp(out, "\0"), 0);
// í (U+00ED)
out = chr_int32(ctx_ptr, 237, &out_len);
EXPECT_EQ(std::string(out, out_len), "\xC3\xAD");

out = chr_int32(ctx_ptr, -158, &out_len);
EXPECT_EQ(std::string(out, out_len), "b");
// Highest 2-byte code point (U+07FF).
out = chr_int64(ctx_ptr, 0x7FF, &out_len);
EXPECT_EQ(std::string(out, out_len), "\xDF\xBF");

out = chr_int64(ctx_ptr, -5, &out_len);
EXPECT_EQ(std::string(out, out_len), "\xFB");
// 3-byte code points.
// Lowest 3-byte code point (U+0800).
out = chr_int32(ctx_ptr, 0x800, &out_len);
EXPECT_EQ(std::string(out, out_len), "\xE0\xA0\x80");

out = chr_int32(ctx_ptr, -340, &out_len);
EXPECT_EQ(std::string(out, out_len), "\xAC");
// € (U+20AC)
out = chr_int32(ctx_ptr, 8364, &out_len);
EXPECT_EQ(std::string(out, out_len), "\xE2\x82\xAC");

out = chr_int64(ctx_ptr, -66, &out_len);
EXPECT_EQ(std::string(out, out_len), "\xBE");
// 日 (U+65E5)
out = chr_int64(ctx_ptr, 26085, &out_len);
EXPECT_EQ(std::string(out, out_len), "\xE6\x97\xA5");

//€
out = chr_int32(ctx_ptr, 128, &out_len);
EXPECT_EQ(std::string(out, out_len), "\x80");
// Highest 3-byte code point (U+FFFF).
out = chr_int32(ctx_ptr, 0xFFFF, &out_len);
EXPECT_EQ(std::string(out, out_len), "\xEF\xBF\xBF");

//œ
out = chr_int64(ctx_ptr, 156, &out_len);
EXPECT_EQ(std::string(out, out_len), "\x9C");
// 4-byte code points.
// Lowest 4-byte code point (U+10000).
out = chr_int64(ctx_ptr, 0x10000, &out_len);
EXPECT_EQ(std::string(out, out_len), "\xF0\x90\x80\x80");

//ÿ
out = chr_int32(ctx_ptr, 255, &out_len);
EXPECT_EQ(std::string(out, out_len), "\xFF");
// 😀 (U+1F600)
out = chr_int64(ctx_ptr, 0x1F600, &out_len);
EXPECT_EQ(std::string(out, out_len), "\xF0\x9F\x98\x80");

// BACKSPACE
out = chr_int64(ctx_ptr, 8, &out_len);
EXPECT_EQ(std::string(out, out_len), "\b");
// Highest valid code point (U+10FFFF).
out = chr_int64(ctx_ptr, 0x10FFFF, &out_len);
EXPECT_EQ(std::string(out, out_len), "\xF4\x8F\xBF\xBF");

// DEVICE CONTROL 3 (DC3)
out = chr_int32(ctx_ptr, 19, &out_len);
EXPECT_EQ(std::string(out, out_len), "\x13");
EXPECT_FALSE(ctx.has_error());

// ESCAPE (ESC)
out = chr_int64(ctx_ptr, 27, &out_len);
EXPECT_EQ(std::string(out, out_len), "\x1B");
// Invalid code points raise an error.
chr_int64(ctx_ptr, -1, &out_len);
EXPECT_EQ(out_len, 0);
EXPECT_TRUE(ctx.get_error().find("not a valid Unicode code point") != std::string::npos)
<< ctx.get_error();
ctx.Reset();

chr_int64(ctx_ptr, 0x110000, &out_len);
EXPECT_EQ(out_len, 0);
EXPECT_TRUE(ctx.get_error().find("not a valid Unicode code point") != std::string::npos)
<< ctx.get_error();
ctx.Reset();

// UTF-16 surrogate range is not a valid code point.
chr_int32(ctx_ptr, 0xD800, &out_len);
EXPECT_EQ(out_len, 0);
EXPECT_TRUE(ctx.get_error().find("not a valid Unicode code point") != std::string::npos)
<< ctx.get_error();
ctx.Reset();
}

TEST(TestStringOps, TestBeginsEnds) {
Expand Down
Loading