fix(cobol): detect fixed-format source that carries sequence numbers

The fixed/free format detector counted a line as fixed-format only when its
first six columns were blank (`^ {6}[ *\-/ ]`). Legacy mainframe COBOL — the
main reason to have a COBOL extractor — puts a sequence NUMBER in columns 1-6,
so every such file scored zero fixed markers and was classified free-format.

In free mode the code is taken verbatim (no column-8 slice), so the sequence
number stays glued to the statement: `000400 MAIN-PARA.` no longer matches the
paragraph anchor `^NAME .? .$`, and the whole procedure layer disappears. A
seven-line program with two paragraphs and a PERFORM collapsed to just the file
and program nodes — zero paragraphs, zero `calls` edges — silently.

Match a sequence area that is six spaces OR six digits, followed by a valid
indicator column (space = code, `*`/`/` = comment, `-` = continuation, `D` =
debug). Once the file is correctly read as fixed-format the existing
column-8..72 slice strips the sequence number and paragraphs/PERFORMs resolve
as they should. Blank-sequence fixed-format and genuinely free-format source are
unaffected (a 6-digit prefix is not something free-format code begins a line
with).

Regression test: a sequence-numbered program now yields its program, both
paragraphs, and the PERFORM `calls` edge — and fails on the old detector.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01JJbLfztSxm2tH5cJwBbe9q

graphify/extractors/cobol.py

@@ -40,8 +40,17 @@ def _code_lines(source: str) -> list[tuple[int, str]]:
     forced_free = bool(
         re.search(r">>\s*SOURCE\s+FORMAT\s+(?:IS\s+)?FREE", source, re.IGNORECASE)
     )
+    # A fixed-format line is a 6-column sequence area followed by the indicator
+    # in column 7. The sequence area is blank on some files but carries a
+    # sequence NUMBER on legacy mainframe source (the historical norm), so match
+    # both: six spaces, or six digits, followed by a valid indicator (space =
+    # code, `*`/`/` = comment, `-` = continuation, `D` = debug). Requiring six
+    # blank columns misclassified every sequence-numbered file as free-format,
+    # which then kept the sequence number in the code and dropped every paragraph
+    # (its `^NAME.$` anchor no longer matched) and PERFORM edge.
     fixed_markers = sum(
-        bool(re.match(r"^ {6}[ *\-/ ]", line)) for line in physical if line.strip()
+        bool(re.match(r"^(?: {6}|\d{6})[ *\-/dD]", line))
+        for line in physical if line.strip()
     )
     nonempty = sum(bool(line.strip()) for line in physical)
     free = forced_free or fixed_markers < max(1, nonempty // 2)

tests/test_cobol_extractor.py

@@ -131,3 +131,28 @@ def test_cobol_malformed_tail_comments_and_strings_do_not_create_phantoms(tmp_pa
     labels = {node["label"].casefold() for node in result["nodes"]}
     assert 'kept' in labels
     assert labels.isdisjoint({'ghost'})
+
+
+def test_cobol_fixed_format_with_sequence_numbers(tmp_path):
+    """Legacy fixed-format source carries a sequence NUMBER in columns 1-6, not
+    blanks. The format detector required six blank columns, so a sequence-numbered
+    file was misread as free-format — the sequence number stayed in the code and
+    every paragraph (its `^NAME.$` anchor no longer matched) and PERFORM edge was
+    lost, leaving only the file and program nodes."""
+    source = tmp_path / "legacy.cob"
+    source.write_text(
+        "000100 IDENTIFICATION DIVISION.\n"
+        "000200 PROGRAM-ID. PAYROLL.\n"
+        "000300 PROCEDURE DIVISION.\n"
+        "000400 MAIN-PARA.\n"
+        "000500     PERFORM INIT-PARA.\n"
+        "000600 INIT-PARA.\n"
+        "000700     DISPLAY 'HELLO'.\n",
+        encoding="utf-8",
+    )
+
+    result = extract([source], cache_root=tmp_path)
+
+    labels = {node["label"] for node in result["nodes"]}
+    assert {"PAYROLL", "MAIN-PARA", "INIT-PARA"} <= labels
+    assert ("MAIN-PARA", "INIT-PARA") in _edge_labels(result, "calls")