ci: Add style checks for some common mistakes

.github/workflows/style.yml

@@ -0,0 +1,21 @@
+name: Style
+
+on:
+  push:
+    branches:
+      - main
+  pull_request:
+    branches:
+      - main
+
+jobs:
+  check_style:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Check out repo
+        uses: actions/checkout@v4
+      - name: Find changed lines
+        id: changed_lines
+        uses: hestonhoffman/changed-lines@v1
+      - name: Verify changed lines
+        run: ./tools/stylecheck.py '${{ steps.changed_lines.outputs.changed_lines }}'

tools/stylecheck.py

@@ -0,0 +1,166 @@
+#!/usr/bin/env python3
+
+import json
+from pathlib import Path
+import re
+import sys
+
+
+MACRO_DEFINE_PATTERN = re.compile(r"^#\s*define\s")
+STRING_LITERAL_PATTERN = re.compile(r"\"[^\"]+?\"")
+
+def function_parameters_are_aligned(match):
+    lines = match.group(1).rstrip().split("\n")
+    if lines[0].endswith(" ="):
+        return True
+
+    if len(lines) < 2:
+        return False
+
+    if lines[1].endswith(" ="):
+        return True
+
+    offset = lines[1].find("(")
+    if offset == -1:
+        return False
+
+    if offset == len(lines[1]) - 1:
+        offset = 3
+
+    expected_num_leading_spaces = offset + 1
+    for line in lines[2:]:
+        num_leading_spaces = len(line) - len(line.lstrip(" "))
+        if num_leading_spaces != expected_num_leading_spaces:
+            return False
+
+    return True
+
+COMMON_MISTAKES = [
+    (
+        "line exceeds 80 columns",
+        re.compile(r"^.{81}()", re.MULTILINE),
+    ),
+    (
+        "missing space before parentheses",
+        re.compile(r"\w()\("),
+        ("unless-line-matches", MACRO_DEFINE_PATTERN),
+        ("unless-found-inside", STRING_LITERAL_PATTERN),
+    ),
+    (
+        "missing space after cast",
+        re.compile(r"\([^()]+\)()\w"),
+    ),
+    (
+        "missing space in pointer declaration",
+        re.compile(r"\w+()\* \w+"),
+    ),
+    (
+        "missing space in pointer declaration",
+        re.compile(r"\w+ \*()\w+"),
+        ("unless-line-matches", re.compile(r"\s+return \*")),
+    ),
+    (
+        "blank line after block start",
+        re.compile("{\n(\n)"),
+    ),
+    (
+        "blank line before block end",
+        re.compile("\n(\n)}"),
+    ),
+    (
+        "two or more consecutive blank lines",
+        re.compile("\n(\n{2})"),
+    ),
+    (
+        "opening brace on the same line as the statement opening it",
+        re.compile(r"^.+\)[^\n]*({)", re.MULTILINE),
+        ("unless-line-matches", MACRO_DEFINE_PATTERN),
+        ("unless-found-inside", STRING_LITERAL_PATTERN),
+    ),
+    (
+        "incorrectly formatted function definition",
+        re.compile(r"^(static [^;{]+){", re.MULTILINE),
+        ("unless-true", function_parameters_are_aligned),
+    ),
+]
+
+COMMENT_PATTERN = re.compile(r"\/\*(.+?)\*\/", re.DOTALL)
+
+EXCLUDED_SOURCES = {
+    "gum/dlmalloc.c",
+    "gum/gummetalhash.c",
+    "gum/gummetalhash.h",
+    "gum/gumprintf.c",
+    "gum/valgrind.h",
+}
+
+
+def main():
+    changed_lines = json.loads(sys.argv[1])
+
+    repo_dir = Path(__file__).parent.parent.resolve()
+    files_to_check = [Path(repo_dir / f) for f in changed_lines.keys()]
+
+    num_mistakes_found = 0
+    for path in files_to_check:
+        relpath = path.relative_to(repo_dir).as_posix()
+        if relpath in EXCLUDED_SOURCES:
+            continue
+
+        code = path.read_text(encoding="utf-8")
+        lines = code.split("\n")
+
+        comment_lines = set()
+        for m in COMMENT_PATTERN.finditer(code):
+            start_offset, end_offset = m.span(1)
+            start_line = offset_to_line(start_offset, code)
+            end_line = offset_to_line(end_offset, code)
+            for i in range(start_line, end_line + 1):
+                comment_lines.add(i)
+
+        for (description, pattern, *predicates) in COMMON_MISTAKES:
+            for match in pattern.finditer(code):
+                match_offset = match.start(1)
+                line_number = offset_to_line(match_offset, code)
+
+                if line_number in comment_lines:
+                    continue
+
+                prev_newline_offset = code.rfind("\n", 0, match_offset)
+                if prev_newline_offset == -1:
+                    prev_newline_offset = 0
+                line_offset = match_offset - prev_newline_offset
+
+                is_actual_mistake = True
+                line = lines[line_number - 1]
+                for (condition, parameter) in predicates:
+                    if condition == "unless-line-matches":
+                        if parameter.match(line) is not None:
+                            is_actual_mistake = False
+                    elif condition == "unless-found-inside":
+                        for m in parameter.finditer(line):
+                            start, end = m.span()
+                            if line_offset >= start and line_offset < end:
+                                is_actual_mistake = False
+                                break
+                    elif condition == "unless-true":
+                        if parameter(match):
+                            is_actual_mistake = False
+                    else:
+                        assert False, "unexpected condition"
+                    if not is_actual_mistake:
+                        break
+
+                if is_actual_mistake and line_number in changed_lines[relpath]:
+                    print(f"{relpath}:{line_number}: {description}")
+                    num_mistakes_found += 1
+
+    sys.exit(0 if num_mistakes_found == 0 else 1)
+
+
+def offset_to_line(i, code):
+    return len(code[:i].split("\n"))
+
+
+if __name__ == "__main__":
+    main()