Added NativeStringType (#2252)

* #include an absolute path didn't work - because paths were taken to always be relative.

* Use TerminatedUnownedStringSlice for literals in output C++.

* Remove Escape/Unescape functions used in slang-token-reader.cpp
Add target type of 'host-cpp' etc to map to the target types.

* Fix some corner cases around string encoding.

* Added unit test for string escaping.
Fixed some assorted escaping bugs.

* Updated test output.

* Added decode test.

* Stop using hex output, to get around 'greedy' aspect. Use octal instead.

2 files changed, 80 insertions, 1 deletions

diff --git a/tools/slang-unit-test/unit-offset-container.cpp b/tools/slang-unit-test/unit-test-offset-container.cpp
index 6a179c319..9d8e3a9ff 100644
--- a/tools/slang-unit-test/unit-offset-container.cpp
+++ b/tools/slang-unit-test/unit-test-offset-container.cpp
@@ -1,4 +1,4 @@
-// unit-test-path.cpp
+// unit-test-offset-container.cpp
 
 #include "../../source/core/slang-offset-container.h"
 
diff --git a/tools/slang-unit-test/unit-test-string-escape.cpp b/tools/slang-unit-test/unit-test-string-escape.cpp
new file mode 100644
index 000000000..337573081
--- /dev/null
+++ b/tools/slang-unit-test/unit-test-string-escape.cpp
@@ -0,0 +1,79 @@
+// unit-test-string-escape.cpp
+
+#include "../../source/core/slang-string-escape-util.h"
+
+#include "tools/unit-test/slang-unit-test.h"
+
+using namespace Slang;
+
+static bool _checkConversion(StringEscapeHandler* handler, const UnownedStringSlice& check)
+{
+	StringBuilder buf;
+	handler->appendEscaped(check, buf);
+
+	StringBuilder decode;
+	handler->appendUnescaped(buf.getUnownedSlice(), decode);
+
+	return decode == check;
+}
+
+static bool _checkDecode(const UnownedStringSlice& encoded, const UnownedStringSlice& decoded)
+{
+	auto handler = StringEscapeUtil::getHandler(StringEscapeUtil::Style::Cpp);
+
+	StringBuilder buf;
+	StringEscapeUtil::appendUnquoted(handler, encoded, buf);
+	return buf == decoded;
+}
+
+#define SLANG_ENCODED_DECODED(x) \
+		const auto encoded = toSlice(#x); \
+		const auto decoded = toSlice(x); 
+
+SLANG_UNIT_TEST(StringEscape)
+{
+	// Check greedy hex digits
+	{
+		// \x can have any number of hex digits
+		const char text[] = "\x000001";
+		SLANG_ASSERT(SLANG_COUNT_OF(text) == 2 && text[0] == 1);
+	}
+
+	// Check octal greedy
+	{
+		//\ + up to 3 octal digits
+		const char text[] = "\0011";
+		SLANG_ASSERT(SLANG_COUNT_OF(text) == 3 && text[0] == 1 && text[1] == '1');
+
+		const char text2[] = "\78";
+		SLANG_ASSERT(SLANG_COUNT_OF(text2) == 3 && text2[0] == 7 && text2[1] == '8');
+	}
+
+	{
+		auto handler = StringEscapeUtil::getHandler(StringEscapeUtil::Style::Cpp);
+
+		SLANG_CHECK(_checkConversion(handler, toSlice("\0\1\2""2")));
+	}
+
+	{
+		auto handler = StringEscapeUtil::getHandler(StringEscapeUtil::Style::Cpp);
+
+		// We can't just use '\uxxxx', because it has to be translatable into an output character in MSVC (not into utf8)
+		// Can make work perhaps with something like
+		// #pragma execution_character_set("utf-8") 
+		// But for now we don't worry
+		// 
+		// Visual Studio does not appear to support '\U' by default, presumably because wchar_t is 16 bits 
+
+		{
+			SLANG_ENCODED_DECODED("\a\b\0hey~\u0023\n\0");
+			SLANG_CHECK(_checkDecode(encoded, decoded));
+		}
+
+		{
+			SLANG_ENCODED_DECODED("\n\v\b\t\1\02\003\x5z\x00007f\0");
+			SLANG_CHECK(_checkDecode(encoded, decoded));
+		}
+	}
+}
+