Merge pull request #9012 from ethereum/wrong-offset-for-utf8

Fix caret position for errors with utf source

liblangutil/SourceReferenceFormatterHuman.cpp

@@ -21,6 +21,7 @@
 #include <liblangutil/SourceReferenceFormatterHuman.h>
 #include <liblangutil/Scanner.h>
 #include <liblangutil/Exceptions.h>
+#include <libsolutil/UTF8.h>
 #include <iomanip>
 
 using namespace std;
@@ -103,12 +104,13 @@ void SourceReferenceFormatterHuman::printSourceLocation(SourceReference const& _
 		m_stream << leftpad << ' ';
 		frameColored() << '|';
 		m_stream << ' ';
+
 		for_each(
 			_ref.text.cbegin(),
-			_ref.text.cbegin() + _ref.startColumn,
+			_ref.text.cbegin() + numCodepoints(_ref.text.substr(0, _ref.startColumn)),
 			[this](char ch) { m_stream << (ch == '\t' ? '\t' : ' '); }
 		);
-		diagColored() << string(locationLength, '^');
+		diagColored() << string(numCodepoints(_ref.text.substr(_ref.startColumn, locationLength)), '^');
 		m_stream << '\n';
 	}
 	else

libsolutil/UTF8.cpp

@@ -138,4 +138,13 @@ bool validateUTF8(std::string const& _input, size_t& _invalidPosition)
 	return validateUTF8(reinterpret_cast<unsigned char const*>(_input.c_str()), _input.length(), _invalidPosition);
 }
 
+size_t numCodepoints(std::string const& _utf8EncodedInput)
+{
+	size_t codepoint = 0;
+	for (char c: _utf8EncodedInput)
+		codepoint += (c & 0xc0) != 0x80;
+
+	return codepoint;
+}
+
 }

libsolutil/UTF8.h

@@ -38,4 +38,6 @@ inline bool validateUTF8(std::string const& _input)
 	return validateUTF8(_input, invalidPos);
 }
 
+size_t numCodepoints(std::string const& _utf8EncodedInput);
+
 }

scripts/isolate_tests.py

@@ -13,7 +13,7 @@ import hashlib
 from os.path import join, isfile
 
 def extract_test_cases(path):
-    lines = open(path, mode='r', encoding='utf8').read().splitlines()
+    lines = open(path, encoding="utf8", errors='ignore', mode='r').read().splitlines()
 
     inside = False
     delimiter = ''

scripts/wasm-rebuild/docker-scripts/isolate_tests.py

@@ -8,7 +8,7 @@ from os.path import join, isfile
 
 
 def extract_test_cases(path):
-    lines = open(path, 'rb').read().splitlines()
+    lines = open(path, encoding="utf8", errors='ignore', mode='rb').read().splitlines()
 
     inside = False
     delimiter = ''

test/cmdlineTests/message_format_utf16/err

@@ -0,0 +1,11 @@
+Warning: SPDX license identifier not provided in source file. Before publishing, consider adding a comment containing "SPDX-License-Identifier: <SPDX-License>" to each source file. Use "SPDX-License-Identifier: UNLICENSED" for non-open-source code. Please see https://spdx.org for more information.
+--> message_format_utf16/input.sol
+
+Warning: Source file does not specify required compiler version!
+--> message_format_utf16/input.sol
+
+Warning: Statement has no effect.
+ --> message_format_utf16/input.sol:2:58:
+  |
+2 | /* ©©©©ﾩ©©©©© 2017 */ constructor () public { "©©©©ﾩ©©©©©" ; }
+  |                                               ^^^^^^^^^^^^

test/cmdlineTests/message_format_utf16/input.sol

@@ -0,0 +1,3 @@
+contract Foo {
+/* ©©©©ﾩ©©©©© 2017 */ constructor () public { "©©©©ﾩ©©©©©" ; }
+}