Find bad comments in one pass, add awareness of TEMPLATE

tools/checkorder/checkorder.py

@@ -5,7 +5,6 @@
 from isledecomp.dir import walk_source_dir
 from isledecomp.parser import find_code_blocks
 from isledecomp.parser.util import (
-    match_offset_comment,
     is_exact_offset_comment
 )
 
@@ -16,16 +15,6 @@ def sig_truncate(sig: str) -> str:
     return f"{sig[:47]}{'...' if len(sig) >= 50 else ''}"
 
 
-def get_inexact_offset_comments(stream: TextIO) -> [tuple]:
-    """Read the file stream and return the line number and string
-       for any offset comments that don't exactly match the template."""
-    return ([
-        (line_no, line.strip())
-        for line_no, line in enumerate(stream)
-        if match_offset_comment(line) and not is_exact_offset_comment(line)
-    ])
-
-
 def check_file(filename: str, verbose: bool = False) -> bool:
     """Open and read the given file, then check whether the code blocks
        are in order. If verbose, print each block."""
@@ -35,7 +24,10 @@ def check_file(filename: str, verbose: bool = False) -> bool:
         # TODO: Should combine these checks if/when we refactor.
         # This is just for simplicity / proof of concept.
         f.seek(os.SEEK_SET, 0)
-        bad_comments = get_inexact_offset_comments(f)
+
+    bad_comments = [(block.start_line, block.offset_comment)
+                    for block in code_blocks
+                    if not is_exact_offset_comment(block.offset_comment)]
 
     just_offsets = [block.offset for block in code_blocks]
     sorted_offsets = sorted(just_offsets)

tools/isledecomp/isledecomp/parser/parser.py

@@ -26,6 +26,7 @@ def find_code_blocks(stream: TextIO) -> List[CodeBlock]:
     blocks = []
 
     offset = None
+    offset_comment = None
     function_sig = None
     start_line = None
     state = ReaderState.WANT_OFFSET
@@ -39,7 +40,8 @@ def find_code_blocks(stream: TextIO) -> List[CodeBlock]:
                 block = CodeBlock(offset=offset,
                                   signature=function_sig,
                                   start_line=start_line,
-                                  end_line=line_no)
+                                  end_line=line_no,
+                                  offset_comment=offset_comment)
 
                 blocks.append(block)
                 state = ReaderState.WANT_OFFSET
@@ -50,7 +52,8 @@ def find_code_blocks(stream: TextIO) -> List[CodeBlock]:
                 block = CodeBlock(offset=offset,
                                   signature=function_sig,
                                   start_line=start_line,
-                                  end_line=line_no - 1)
+                                  end_line=line_no - 1,
+                                  offset_comment=offset_comment)
 
                 blocks.append(block)
                 state = ReaderState.WANT_OFFSET
@@ -67,6 +70,7 @@ def find_code_blocks(stream: TextIO) -> List[CodeBlock]:
             match = match_offset_comment(line)
             if match is not None:
                 offset = int(match, 16)
+                offset_comment = line.strip()
                 start_line = line_no
                 state = ReaderState.WANT_SIG
 

tools/isledecomp/isledecomp/parser/util.py

@@ -4,20 +4,18 @@
 
 
 CodeBlock = namedtuple('CodeBlock',
-                       ['offset', 'signature', 'start_line', 'end_line'])
+                       ['offset', 'signature', 'start_line', 'end_line',
-
+                        'offset_comment'])
-
-FunctionOffset = namedtuple('FunctionOffset',
-                            ['raw', 'address', 'is_stub'])
 
 
 # To match a reasonable variance of formatting for the offset comment
-offsetCommentRegex = re.compile(r'//\s*OFFSET:\s*\w+\s+(?:0x)?([a-f0-9]+)',
+offsetCommentRegex = re.compile(r'//\s*OFFSET:\s*\w+\s+(?:0x)?([a-f0-9]+)(\s+STUB)?',
                                 flags=re.I)
 
 # To match the exact syntax (text upper case, hex lower case, with spaces)
 # that is used in most places
-offsetCommentExactRegex = re.compile(r'^// OFFSET: [A-Z0-9]+ (0x[a-f0-9]+)(?: STUB)?$')
+# TODO: template and stub mutually exclusive?
+offsetCommentExactRegex = re.compile(r'^// OFFSET: [A-Z0-9]+ (0x[a-f0-9]+)(?: STUB| TEMPLATE)?$')
 
 
 def is_blank_or_comment(line: str) -> bool: