Swift: prevent accidental revert of modified stub

If one modifies a QL stub but forgets to remove the `// generated`
header comment, codegen will now abort with an error rather than
silently reverting the change.

This is based on the rough heuristic of just counting the lines. If any
change is done to the stub class, the number of lines is bound to be
5 or more.

Author: redsun82

swift/codegen/generators/qlgen.py

@@ -12,7 +12,16 @@
 log = logging.getLogger(__name__)
 
 
-class FormatError(Exception):
+class Error(Exception):
+    def __str__(self):
+        return self.args[0]
+
+
+class FormatError(Error):
+    pass
+
+
+class ModifiedStubMarkedAsGeneratedError(Error):
     pass
 
 
@@ -84,11 +93,22 @@ def get_classes_used_by(cls: ql.Class):
     return sorted(set(t for t in get_types_used_by(cls) if t[0].isupper()))
 
 
-def is_generated(file):
+def _is_generated_stub(file, check_modification=False):
     with open(file) as contents:
         for line in contents:
-            return line.startswith("// generated")
-        return False
+            if not line.startswith("// generated"):
+                return False
+            break
+        else:
+            # no lines
+            return False
+        if check_modification:
+            # one line already read, if we can read 4 other we are past the normal stub generation
+            line_threshold = 4
+            if sum(1 for _ in zip(range(line_threshold), contents)) == line_threshold:
+                raise ModifiedStubMarkedAsGeneratedError(
+                    f"{file.name} stub was modified but is still marked as generated")
+        return True
 
 
 def format(codeql, files):
@@ -138,7 +158,7 @@ def generate(opts, renderer):
     test_out = opts.ql_test_output
     missing_test_source_filename = "MISSING_SOURCE.txt"
     existing = {q for q in out.rglob("*.qll")}
-    existing |= {q for q in stub_out.rglob("*.qll") if is_generated(q)}
+    existing |= {q for q in stub_out.rglob("*.qll") if _is_generated_stub(q, check_modification=True)}
     existing |= {q for q in test_out.rglob("*.ql")}
     existing |= {q for q in test_out.rglob(missing_test_source_filename)}
 
@@ -157,7 +177,7 @@ def generate(opts, renderer):
         c.imports = [imports[t] for t in get_classes_used_by(c)]
         renderer.render(c, qll)
         stub_file = stub_out / c.path.with_suffix(".qll")
-        if not stub_file.is_file() or is_generated(stub_file):
+        if not stub_file.is_file() or _is_generated_stub(stub_file):
             stub = ql.Stub(
                 name=c.name, base_import=get_import(qll, opts.swift_dir))
             renderer.render(stub, stub_file)

swift/codegen/test/test_qlgen.py

@@ -2,6 +2,8 @@
 import subprocess
 import sys
 
+import pytest
+
 from swift.codegen.generators import qlgen
 from swift.codegen.lib import ql
 from swift.codegen.test.utils import *
@@ -36,17 +38,22 @@ def children_file(): return ql_output_path() / "GetImmediateParent.qll"
 
 
 @pytest.fixture
-def generate(input, opts, renderer):
+def qlgen_opts(opts):
     opts.ql_stub_output = stub_path()
     opts.ql_output = ql_output_path()
     opts.ql_test_output = ql_test_output_path()
     opts.ql_format = True
     opts.swift_dir = paths.swift_dir
+    return opts
+
+
+@pytest.fixture
+def generate(input, qlgen_opts, renderer):
     renderer.written = []
 
     def func(classes):
         input.classes = classes
-        return run_generation(qlgen.generate, opts, renderer)
+        return run_generation(qlgen.generate, qlgen_opts, renderer)
 
     return func
 
@@ -349,7 +356,7 @@ def test_empty_cleanup(generate, renderer):
     assert renderer.mock_calls[-1] == mock.call.cleanup(set())
 
 
-def test_non_empty_cleanup(opts, generate, renderer, tmp_path):
+def test_non_empty_cleanup(opts, generate, renderer):
     ql_a = opts.ql_output / "A.qll"
     ql_b = opts.ql_output / "B.qll"
     stub_a = opts.ql_stub_output / "A.qll"
@@ -369,6 +376,13 @@ def test_non_empty_cleanup(opts, generate, renderer, tmp_path):
         {ql_a, ql_b, stub_a, test_a, test_b})
 
 
+def test_modified_stub_still_generated(qlgen_opts, renderer):
+    stub = qlgen_opts.ql_stub_output / "A.qll"
+    write(stub, "// generated\n\n\n\nsomething\n")
+    with pytest.raises(qlgen.ModifiedStubMarkedAsGeneratedError):
+        run_generation(qlgen.generate, qlgen_opts, renderer)
+
+
 def test_test_missing_source(generate_tests):
     generate_tests([
         schema.Class("A"),