utun2/c_util

#!/usr/bin/env python3
"""
c_util - C code navigation utility
Commands:
  toc                     - Show table of contents for all files
  description <names...>  - Show selected items with comments
  show <name1> [name2]    - Show code for functions/structs/enums
"""

import sys
import re

C_KEYWORDS = {
    'auto', 'break', 'case', 'char', 'const', 'continue', 'default', 'do',
    'double', 'else', 'enum', 'extern', 'float', 'for', 'goto', 'if',
    'int', 'long', 'register', 'return', 'short', 'signed', 'sizeof', 'static',
    'struct', 'switch', 'typedef', 'union', 'unsigned', 'void', 'volatile', 'while',
    '_Alignas', '_Alignof', '_Atomic', '_Bool', '_Complex', '_Generic', '_Imaginary',
    '_Noreturn', '_Static_assert', '_Thread_local'
}

CONTROL_FLOW = {'if', 'for', 'while', 'switch', 'return', 'else'}


def read_filelist(filename='filelist.txt'):
    """Read list of files to process."""
    with open(filename, 'r') as f:
        return [line.strip() for line in f if line.strip()]


def load_files(files):
    """Load contents of all files."""
    contents = {}
    for f in files:
        try:
            with open(f, 'r') as fd:
                contents[f] = fd.read().splitlines()
        except Exception as e:
            print(f"Warning: could not read {f}: {e}", file=sys.stderr)
    return contents


def preprocess_lines(lines):
    """
    Preprocess lines:
    - Remove preprocessor directives (#...)
    - Remove comments (/* */ and //)
    - Track pre_lines (comments before declarations)
    Returns list of tuples: (original_line, is_code, pre_lines)
    """
    result = []
    multi_comment = False
    pre_lines = []
    
    for line in lines:
        stripped = line.strip()
        
        # Handle multi-line comments
        if multi_comment:
            if '*/' in line:
                multi_comment = False
            pre_lines.append(line)
            result.append((line, False, []))
            continue
        
        # Start of multi-line comment
        if '/*' in line:
            end_pos = line.find('*/')
            if end_pos == -1:
                # Comment continues to next line
                multi_comment = True
                pre_lines.append(line)
                result.append((line, False, []))
                continue
            else:
                # Single-line block comment
                pre_lines.append(line)
                result.append((line, False, []))
                continue
        
        # Single-line comment
        if stripped.startswith('//'):
            pre_lines.append(line)
            result.append((line, False, []))
            continue
        
        # Preprocessor directive - remove but don't add to pre_lines
        if stripped.startswith('#'):
            result.append((line, False, []))
            continue
        
        # Empty line - keep as separator but not code
        if not stripped:
            result.append((line, False, []))
            continue
        
        # This is actual code
        current_pre = pre_lines[:]
        pre_lines = []
        result.append((line, True, current_pre))
    
    return result


def compute_brace_levels(lines_info):
    """
    Compute brace nesting level for each line.
    Returns list of levels (level before processing the line).
    """
    levels = []
    brace_level = 0
    
    for line, is_code, _ in lines_info:
        levels.append(brace_level)
        
        if not is_code:
            continue
        
        stripped = line.strip()
        
        # Skip preprocessor (already filtered but double-check)
        if stripped.startswith('#'):
            continue
        
        # Skip extern "C" { blocks (C++ linkage)
        if 'extern "C"' in stripped:
            # Count braces in extern "C" line separately
            if '{' in stripped:
                # Don't count this brace as it opens/closes C linkage block
                pass
            continue
        
        # Skip do { ... } while(0) - not real nesting
        if stripped.startswith('do {'):
            continue
        if stripped.startswith('}') and 'while(0)' in stripped:
            continue
        
        # Count braces
        brace_level += line.count('{') - line.count('}')
    
    return levels


def extract_identifier_before_paren(line):
    """
    Extract function name before first '('.
    Returns (name, return_type) or (None, None) if invalid.
    """
    paren_pos = line.find('(')
    if paren_pos <= 0:
        return None, None
    
    before = line[:paren_pos].strip()
    
    # Split into parts (handle pointers, qualifiers, etc.)
    # e.g., "static int * func_name" -> ["static", "int", "*", "func_name"]
    parts = before.split()
    
    if not parts:
        return None, None
    
    # Last part should be function name
    name = parts[-1].strip('*').strip()
    
    # Check if it's a valid identifier
    if not name.isidentifier():
        return None, None
    
    # Check if it's a keyword
    if name in C_KEYWORDS or name in CONTROL_FLOW:
        return None, None
    
    # Return type is everything except name
    ret_type = ' '.join(parts[:-1]) if len(parts) > 1 else 'void'
    
    return name, ret_type


def find_matching_brace(lines_info, start_idx, start_level):
    """
    Find line index of matching closing brace.
    Starts from start_idx, looking for level to return to start_level.
    """
    level = start_level
    
    for i in range(start_idx, len(lines_info)):
        line, is_code, _ = lines_info[i]
        
        if not is_code:
            continue
        
        stripped = line.strip()
        
        # Skip preprocessor
        if stripped.startswith('#'):
            continue
        
        # Skip do-while
        if stripped.startswith('do {'):
            continue
        if stripped.startswith('}') and 'while(0)' in stripped:
            continue
        
        # Check current level before this line
        if level == start_level and i > start_idx and '}' in line:
            return i
        
        # Update level
        level += line.count('{') - line.count('}')
    
    return len(lines_info) - 1


def extract_declarations(lines):
    """
    Extract all function, struct, and enum declarations.
    Returns (functions, structs, enums).
    """
    functions = []
    structs = []
    enums = []
    
    # Preprocess
    lines_info = preprocess_lines(lines)
    
    # Compute brace levels
    brace_levels = compute_brace_levels(lines_info)
    
    i = 0
    n = len(lines_info)
    
    while i < n:
        line, is_code, pre_lines = lines_info[i]
        
        if not is_code:
            i += 1
            continue
        
        stripped = line.strip()
        level = brace_levels[i]
        
        # Only process declarations at level 0
        if level != 0:
            i += 1
            continue
        
        # Check for typedef struct/enum/union
        if stripped.startswith('typedef struct'):
            name, decl_end = parse_typedef_struct(lines_info, i)
            if name:
                structs.append({
                    'name': name,
                    'pre': pre_lines,
                    'start': i,
                    'end': decl_end,
                    'line_count': decl_end - i + 1
                })
                i = decl_end + 1
                continue
        
        if stripped.startswith('typedef enum'):
            name, decl_end = parse_typedef_enum(lines_info, i)
            if name:
                enums.append({
                    'name': name,
                    'pre': pre_lines,
                    'start': i,
                    'end': decl_end,
                    'line_count': decl_end - i + 1
                })
                i = decl_end + 1
                continue
        
        # Check for simple struct/enum declarations
        if stripped.startswith('struct ') and '{' in stripped:
            name, decl_end = parse_simple_struct(lines_info, i)
            if name:
                structs.append({
                    'name': name,
                    'pre': pre_lines,
                    'start': i,
                    'end': decl_end,
                    'line_count': decl_end - i + 1
                })
                i = decl_end + 1
                continue
        
        if stripped.startswith('enum ') and '{' in stripped:
            name, decl_end = parse_simple_enum(lines_info, i)
            if name:
                enums.append({
                    'name': name,
                    'pre': pre_lines,
                    'start': i,
                    'end': decl_end,
                    'line_count': decl_end - i + 1
                })
                i = decl_end + 1
                continue
        
        # Check for function declarations
        if '(' in stripped:
            func_info = parse_function(lines_info, i)
            if func_info:
                functions.append(func_info)
                i = func_info['end'] + 1
                continue
        
        i += 1
    
    return functions, structs, enums


def parse_typedef_struct(lines_info, start_idx):
    """Parse typedef struct { ... } name;"""
    # Collect lines until we have complete declaration
    decl_lines = []
    brace_count = 0
    
    for i in range(start_idx, len(lines_info)):
        line, is_code, _ = lines_info[i]
        decl_lines.append(line)
        
        if not is_code:
            continue
        
        stripped = line.strip()
        
        # Count braces (skip do-while)
        if not stripped.startswith('do {'):
            if not (stripped.startswith('}') and 'while(0)' in stripped):
                brace_count += line.count('{') - line.count('}')
        
        # If we've closed all braces and found semicolon, we're done
        if brace_count == 0 and stripped.endswith(';'):
            # Extract name from "} name;"
            decl_text = ' '.join([l.strip() for l in decl_lines])
            close_brace = decl_text.rfind('}')
            semi = decl_text.find(';', close_brace)
            
            if close_brace > 0 and semi > close_brace:
                name = decl_text[close_brace + 1:semi].strip()
                return name, i
            return None, i
    
    return None, len(lines_info) - 1


def parse_typedef_enum(lines_info, start_idx):
    """Parse typedef enum { ... } name;"""
    # Same logic as typedef struct
    decl_lines = []
    brace_count = 0
    
    for i in range(start_idx, len(lines_info)):
        line, is_code, _ = lines_info[i]
        decl_lines.append(line)
        
        if not is_code:
            continue
        
        stripped = line.strip()
        
        if not stripped.startswith('do {'):
            if not (stripped.startswith('}') and 'while(0)' in stripped):
                brace_count += line.count('{') - line.count('}')
        
        if brace_count == 0 and stripped.endswith(';'):
            decl_text = ' '.join([l.strip() for l in decl_lines])
            close_brace = decl_text.rfind('}')
            semi = decl_text.find(';', close_brace)
            
            if close_brace > 0 and semi > close_brace:
                name = decl_text[close_brace + 1:semi].strip()
                return name, i
            return None, i
    
    return None, len(lines_info) - 1


def parse_simple_struct(lines_info, start_idx):
    """Parse struct name { ... };"""
    line, is_code, _ = lines_info[start_idx]
    stripped = line.strip()
    
    # Extract name: "struct name {"
    match = re.match(r'struct\s+(\w+)\s*\{', stripped)
    if not match:
        return None, start_idx
    
    name = match.group(1)
    
    # Find closing brace
    brace_count = 1
    for i in range(start_idx + 1, len(lines_info)):
        line, is_code, _ = lines_info[i]
        
        if not is_code:
            continue
        
        stripped = line.strip()
        
        if not stripped.startswith('do {'):
            if not (stripped.startswith('}') and 'while(0)' in stripped):
                brace_count += line.count('{') - line.count('}')
        
        if brace_count == 0:
            return name, i
    
    return name, len(lines_info) - 1


def parse_simple_enum(lines_info, start_idx):
    """Parse enum name { ... };"""
    line, is_code, _ = lines_info[start_idx]
    stripped = line.strip()
    
    # Extract name: "enum name {"
    match = re.match(r'enum\s+(\w+)\s*\{', stripped)
    if not match:
        return None, start_idx
    
    name = match.group(1)
    
    # Find closing brace
    brace_count = 1
    for i in range(start_idx + 1, len(lines_info)):
        line, is_code, _ = lines_info[i]
        
        if not is_code:
            continue
        
        stripped = line.strip()
        
        if not stripped.startswith('do {'):
            if not (stripped.startswith('}') and 'while(0)' in stripped):
                brace_count += line.count('{') - line.count('}')
        
        if brace_count == 0:
            return name, i
    
    return name, len(lines_info) - 1


def parse_function(lines_info, start_idx):
    """Parse function declaration at level 0."""
    line, is_code, pre_lines = lines_info[start_idx]
    stripped = line.strip()
    
    # Extract function name
    name, ret_type = extract_identifier_before_paren(stripped)
    
    if not name:
        return None
    
    # Check if this is a forward declaration (ends with ;)
    if stripped.endswith(';'):
        return None
    
    # Find opening brace - function body must start with {
    # Check current line and next line only
    decl_end = start_idx
    found_brace = False
    
    # Check current line first - look for { after )
    paren_end = stripped.rfind(')')
    if paren_end > 0:
        after_paren = stripped[paren_end:]
        if '{' in after_paren:
            decl_end = start_idx
            found_brace = True
    
    # If not found, check next few lines (up to 3) for {
    # But stop if we encounter ; (end of prototype)
    if not found_brace:
        for offset in range(1, 4):  # Check next 3 lines
            if start_idx + offset >= len(lines_info):
                break
            
            check_line, check_is_code, _ = lines_info[start_idx + offset]
            if not check_is_code:
                continue
            
            check_stripped = check_line.strip()
            
            # If we hit a semicolon, this is a prototype, not a definition
            if check_stripped.endswith(';'):
                return None
            
            # If we found opening brace, this is the function body
            # Check if '{' is present in the line (not necessarily at start due to formatting)
            if '{' in check_line:
                decl_end = start_idx + offset
                found_brace = True
                break
    
    if not found_brace:
        # No body - probably forward declaration or macro, skip
        return None
    
    # Find closing brace
    closing = find_matching_brace(lines_info, decl_end + 1, 0)
    
    # Extract arguments - need to collect all lines from start to decl_end
    # to handle multi-line function signatures
    sig_lines = []
    for idx in range(start_idx, decl_end + 1):
        l, is_code, _ = lines_info[idx]
        if is_code:
            sig_lines.append(l)
    
    sig_text = ' '.join(sig_lines)
    paren_start = sig_text.find('(')
    paren_end = sig_text.rfind(')')
    if paren_start > 0 and paren_end > paren_start:
        args = sig_text[paren_start + 1:paren_end].strip()
    else:
        args = ''
    
    return {
        'type': 'function',
        'name': name,
        'args': args,
        'ret': ret_type,
        'pre': pre_lines,
        'start': start_idx,
        'end': closing,
        'decl_end': decl_end,  # Store the line with opening brace
        'line_count': closing - start_idx + 1
    }


def line_checksum(line):
    """Calculate checksum: sum of all character codes modulo 256, as 2 hex digits."""
    checksum = sum(ord(c) for c in line) % 256
    return f"{checksum:02X}"


def normalize_whitespace(text):
    """Normalize whitespace: replace multiple spaces/newlines with single space."""
    # Replace all whitespace (spaces, tabs, newlines) with single space
    text = re.sub(r'\s+', ' ', text)
    # Strip leading/trailing whitespace
    return text.strip()


def get_function_signature(lines, func):
    """Get normalized function signature (like prototype)."""
    # Get all lines from start to the line with opening brace
    start = func['start']
    end = func.get('decl_end', func['start'])  # We need to track decl_end in parse_function
    
    # Collect signature lines
    sig_lines = []
    for i in range(start, min(end + 1, len(lines))):
        line = lines[i]
        # Remove comments
        if '//' in line:
            line = line[:line.index('//')]
        sig_lines.append(line)
    
    # Join and normalize
    sig = ' '.join(sig_lines)
    sig = normalize_whitespace(sig)
    
    # Remove the opening brace if present
    if '{' in sig:
        sig = sig[:sig.index('{')].strip()
    
    # Ensure it ends with semicolon
    if not sig.endswith(';'):
        sig += ';'
    
    return sig


def get_declaration_text(lines, decl, include_comments=False):
    """Get normalized declaration text for struct/enum."""
    start = decl['start']
    end = decl['end']
    
    # Collect all lines
    decl_lines = []
    for i in range(start, min(end + 1, len(lines))):
        line = lines[i]
        if not include_comments:
            # Remove comments
            if '//' in line:
                line = line[:line.index('//')]
        decl_lines.append(line)
    
    if include_comments:
        # Normalize: one tab indent, remove empty lines, normalize whitespace
        result_lines = []
        for line in decl_lines:
            # Skip empty lines
            if not line.strip():
                continue
            # Normalize whitespace and use one tab for indentation
            normalized = normalize_whitespace(line)
            if normalized:
                result_lines.append('\t' + normalized)
        return '\n'.join(result_lines)
    else:
        # Join and normalize for toc mode
        text = ' '.join(decl_lines)
        text = normalize_whitespace(text)
        return text


def get_function_comments(pre_lines):
    """Extract function comments from pre_lines.
    
    Returns:
        - Last /* */ block if found (including single-line /* */)
        - Or consecutive // lines (without empty lines or non-comment lines)
        - Empty string if no suitable comments
    """
    if not pre_lines:
        return ""
    
    # Look for last /* */ block
    last_block_comment = ""
    in_block = False
    block_start = -1
    
    for i, line in enumerate(pre_lines):
        stripped = line.strip()
        if stripped.startswith('/*'):
            # Check if it's a single-line comment /* ... */
            if stripped.endswith('*/'):
                # Single-line block comment
                last_block_comment = line
            else:
                # Start of multi-line block comment
                in_block = True
                block_start = i
        elif in_block and stripped.endswith('*/'):
            in_block = False
            # Extract the block
            block_lines = pre_lines[block_start:i+1]
            last_block_comment = '\n'.join(block_lines)
        elif stripped.startswith('*/'):
            in_block = False
    
    if last_block_comment:
        return last_block_comment
    
    # Look for consecutive // lines (from the end, backwards)
    comment_lines = []
    for line in reversed(pre_lines):
        stripped = line.strip()
        if stripped.startswith('//'):
            comment_lines.insert(0, stripped)
        elif not stripped:  # Skip empty lines
            continue
        else:  # Non-comment line breaks the sequence
            break
    
    return '\n'.join(comment_lines) if comment_lines else ""


def main():
    if len(sys.argv) < 2:
        print("Usage: c_util toc | description <name1> [name2] ... | show <name1> [name2] ...")
        sys.exit(1)
    
    cmd = sys.argv[1]
    
    files = read_filelist()
    contents = load_files(files)
    
    # Parse all files
    project_functions = {}
    project_structs = {}
    project_enums = {}
    
    for f, lines in contents.items():
        funcs, strs, enums = extract_declarations(lines)
        project_functions[f] = funcs
        project_structs[f] = strs
        project_enums[f] = enums
    
    if cmd == 'toc':
        first_file = True
        for f in files:
            if f not in contents:
                continue
            
            # Get all declarations for this file
            declarations = []
            
            # Add functions
            for func in project_functions.get(f, []):
                sig = get_function_signature(contents[f], func)
                declarations.append((func['start'], func['end'], sig))
            
            # Add structs
            for s in project_structs.get(f, []):
                text = get_declaration_text(contents[f], s)
                declarations.append((s['start'], s['end'], text))
            
            # Add enums
            for e in project_enums.get(f, []):
                text = get_declaration_text(contents[f], e)
                declarations.append((e['start'], e['end'], text))
            
            if not declarations:
                continue
            
            # Sort by line number
            declarations.sort(key=lambda x: x[0])
            
            # Print empty line between files (except before first)
            if not first_file:
                print()
            first_file = False
            
            # Print filename header
            print(f"{f}:")
            
            # Print in new format
            for start, end, text in declarations:
                print(f"[{start + 1}-{end + 1}] {text}")
    
    elif cmd == 'description':
        # Get list of names to look for
        if len(sys.argv) < 3:
            print("Usage: c_util description <name1> [name2] ...")
            sys.exit(1)
        
        target_names = set(sys.argv[2:])
        first_file = True
        
        for f in files:
            if f not in contents:
                continue
            
            # Collect all matching declarations with full info
            declarations = []
            
            # Check functions
            for func in project_functions.get(f, []):
                if func['name'] in target_names:
                    # Get signature
                    sig = get_function_signature(contents[f], func)
                    # Get comments
                    comments = get_function_comments(func['pre'])
                    declarations.append((func['start'], func['end'], 'function', sig, comments))
            
            # Check structs
            for s in project_structs.get(f, []):
                if s['name'] in target_names:
                    # Get full text with formatting and comments
                    text = get_declaration_text(contents[f], s, include_comments=True)
                    declarations.append((s['start'], s['end'], 'struct', s['name'], text))
            
            # Check enums
            for e in project_enums.get(f, []):
                if e['name'] in target_names:
                    # Get full text with formatting and comments
                    text = get_declaration_text(contents[f], e, include_comments=True)
                    declarations.append((e['start'], e['end'], 'enum', e['name'], text))
            
            if not declarations:
                continue
            
            # Sort by line number
            declarations.sort(key=lambda x: x[0])
            
            # Print empty line between files (except before first)
            if not first_file:
                print()
            first_file = False
            
            # Print filename header
            print(f"{f}:")
            
            # Print declarations with full formatting
            for start, end, decl_type, content, extra in declarations:
                if decl_type == 'function':
                    # Print comments first if any (without line numbers)
                    if extra:
                        print(extra)
                    # Print function signature with line number and checksum
                    lines = contents[f]
                    for line_idx in range(start, end + 1):
                        if line_idx < len(lines):
                            line = lines[line_idx]
                            line_num = line_idx + 1
                            checksum = line_checksum(line)
                            print(f"{line_num} {checksum}: {line}")
                else:
                    # struct or enum - print with line numbers and checksums
                    lines = contents[f]
                    for line_idx in range(start, end + 1):
                        if line_idx < len(lines):
                            line = lines[line_idx]
                            line_num = line_idx + 1
                            checksum = line_checksum(line)
                            print(f"{line_num} {checksum}: {line}")
    
    elif cmd == 'show':
        if len(sys.argv) < 3:
            print("Usage: c_util show <name1> [name2] ...")
            sys.exit(1)
        
        target_names = set(sys.argv[2:])
        first_file = True
        found_any = False
        
        for f in files:
            if f not in contents:
                continue
            
            # Collect all matching items
            items = []
            
            # Check functions
            for func in project_functions.get(f, []):
                if func['name'] in target_names:
                    items.append(('function', func))
                    found_any = True
            
            # Check structs
            for s in project_structs.get(f, []):
                if s['name'] in target_names:
                    items.append(('struct', s))
                    found_any = True
            
            # Check enums
            for e in project_enums.get(f, []):
                if e['name'] in target_names:
                    items.append(('enum', e))
                    found_any = True
            
            if not items:
                continue
            
            # Sort by line number
            items.sort(key=lambda x: x[1]['start'])
            
            # Print empty line between files (except before first)
            if not first_file:
                print()
            first_file = False
            
            # Print filename header
            print(f"{f}:")
            
            # Print items with line numbers and checksums
            for item_type, item in items:
                lines = contents[f]
                start = item['start']
                end = item['end']
                
                # Print pre lines (comments) without line numbers
                pre = item['pre']
                if pre:
                    for pre_line in pre:
                        print(pre_line)
                
                # Print body with line numbers and checksums
                for line_idx in range(start, end + 1):
                    if line_idx < len(lines):
                        line = lines[line_idx]
                        line_num = line_idx + 1
                        checksum = line_checksum(line)
                        print(f"{line_num} {checksum}: {line}")
        
        if not found_any:
            print("Not found")
            sys.exit(1)
    
    else:
        print("Unknown command")
        sys.exit(1)


if __name__ == "__main__":
    main()