Convert byte column offsets to characters

linter.py

@@ -16,7 +16,25 @@
 )
 
 
-class Cppcheck(Linter):
+def byte_offset_to_index(text, offset):
+    """Return the index in `text` of the character at UTF-8 byte `offset`."""
+    return len(text.encode('utf-8')[:offset].decode('utf-8', 'ignore'))
+
+
+class CppcheckBase(Linter):
+    __abstract__ = True
+
+    def reposition_match(self, line, col, m, vv):
+        if col is not None:
+            # cppcheck reports the column as a byte offset into the UTF-8 line,
+            # so every non-ASCII character before the error shifts it. Sublime
+            # counts characters.
+            col = byte_offset_to_index(vv.select_line(line), col)
+
+        return super().reposition_match(line, col, m, vv)
+
+
+class Cppcheck(CppcheckBase):
     cmd = CMD
     regex = REGEX
     error_stream = util.STREAM_BOTH  # linting errors are on stderr, exceptions like "file not found" on stdout
@@ -31,7 +49,7 @@ class Cppcheck(Linter):
     }
 
 
-class CppcheckPlus(Linter):
+class CppcheckPlus(CppcheckBase):
     cmd = CMD
     regex = REGEX
     error_stream = util.STREAM_BOTH  # linting errors are on stderr, exceptions like "file not found" on stdout

tests/test_columns.py

@@ -0,0 +1,70 @@
+import importlib
+import unittest
+
+import sublime
+from SublimeLinter.lint.linter import VirtualView
+
+
+LinterModule = importlib.import_module('SublimeLinter-cppcheck.linter')
+
+
+# Real output of `cppcheck --template={file}:{line}:{column}:{severity}:{id}:{message}
+# --inline-suppr --quiet --enable=style --language=c m.c` (cppcheck 2.21.0) for this source.
+# cppcheck counts the column in bytes of the UTF-8 line: the 'e' with acute accent is two
+# bytes and the emoji (U+1F600) four.
+SOURCE = (
+    '#include <stdio.h>\n'
+    'int main(void) {\n'
+    '\tchar buf[4];\n'
+    '\tprintf("é\U0001f600"); buf[10] = 0;\n'
+    '\tint x = 5;\n'
+    '\tint *p = 0; *p = 1;\n'
+    '\treturn 0;\n'
+    '}\n'
+)
+OUTPUT = (
+    "m.c:4:23:error:arrayIndexOutOfBounds:Array 'buf[4]' accessed at index 10, which is out of bounds.\n"
+    "m.c:6:15:error:nullPointer:Null pointer dereference: p\n"
+    "m.c:4:28:style:unreadVariable:Variable 'buf[10]' is assigned a value that is never used.\n"
+    "m.c:5:8:style:unreadVariable:Variable 'x' is assigned a value that is never used.\n"
+)
+
+
+class TestColumns(unittest.TestCase):
+    def resolve(self, linter_class):
+        linter = linter_class(sublime.View(0), {})
+        vv = VirtualView(SOURCE)
+        return [
+            (m['line'], linter.reposition_match(m['line'], m['col'], m, vv)[1])
+            for m in linter.find_errors(OUTPUT)
+        ]
+
+    def expected(self):
+        lines = SOURCE.split('\n')
+        return [
+            (3, lines[3].index('[10]')),        # `buf[10]`: cppcheck points at the `[`
+            (5, lines[5].index('p = 1')),       # the dereferenced `p`
+            (3, lines[3].index('= 0')),         # the assignment of `buf[10]`
+            (4, lines[4].index('= 5')),         # `int x = 5`
+        ]
+
+    def test_columns_after_non_ascii_characters_are_characters(self):
+        for linter_class in (LinterModule.Cppcheck, LinterModule.CppcheckPlus):
+            self.assertEqual(self.resolve(linter_class), self.expected())
+
+    def test_ascii_line_is_unchanged(self):
+        f = LinterModule.byte_offset_to_index
+        self.assertEqual(f('\tint x = 5;', 7), 7)
+        self.assertEqual(f('abc', 0), 0)
+
+    def test_multibyte_characters_count_their_bytes(self):
+        f = LinterModule.byte_offset_to_index
+        text = 'é\U0001f600x'        # 2 + 4 + 1 bytes
+        self.assertEqual(f(text, 2), 1)    # after the first character
+        self.assertEqual(f(text, 6), 2)    # `x`
+        self.assertEqual(f(text, 7), 3)    # end of line
+
+    def test_offset_inside_a_character_or_past_the_end_does_not_fail(self):
+        f = LinterModule.byte_offset_to_index
+        self.assertEqual(f('\U0001f600x', 2), 0)
+        self.assertEqual(f('ab', 9), 2)