copyright: a short header is named as one, not as a missing notice

tests/copyright_check_untracked_test.py

@@ -95,6 +95,54 @@ def git(repo, *args):
     rc = run(repo, "--check")
     check(rc.returncode == 0, "and the check is green afterwards", rc.stdout + rc.stderr)
 
+    # engine#2175: a SHORT header (Copyright + SPDX, no licence block) is not a "missing copyright notice".
+    # The message has to name the shape, or a reader hunts for a notice that is on line 1 -- measured on
+    # #2175/#2173 whose audio headers kept only the two lines while #2133 carried the full block.
+    short = NOTICE.splitlines()[:2]
+    with open(os.path.join(repo, "short.h"), "w") as f:
+        f.write("\n".join(short) + "\n// short.h -- the file's own description\n")
+    git(repo, "add", "short.h")
+    rc = run(repo, "--check")
+    check(rc.returncode != 0, "a SHORT header still fails (the licence block is required)", rc.stdout)
+    check("copyright + SPDX present but the licence block is not" in rc.stdout,
+          "and the failure names the SHAPE, not a missing notice", rc.stdout + rc.stderr)
+    check("missing copyright notice: short.h" not in rc.stdout,
+          "and does NOT claim the notice is missing", rc.stdout + rc.stderr)
+    # engine#2175: --fix must REPLACE the stub, not stack the block above it. Before this it left both, so the
+    # file came out with the Copyright/SPDX pair twice, the duplicate rule fired, and --check stayed red -- the
+    # tool's own remedy could not fix the shape it had just named.
+    rc = run(repo, "--fix")
+    check(rc.returncode == 0, "--fix repairs a short header", rc.stdout + rc.stderr)
+    got = open(os.path.join(repo, "short.h")).read()
+    check(got.count("SPDX-License-Identifier") == 1,
+          "and exactly one SPDX line survives the repair", f"{got.count('SPDX-License-Identifier')} found")
+    check("Licensed under the Apache License, Version 2.0" in got,
+          "and the full licence block is there", got[:200])
+    check("the file's own description" in got, "and the file's own header text is kept", got)
+    rc = run(repo, "--check")
+    check(rc.returncode == 0, "and --check is green after the repair", rc.stdout + rc.stderr)
+    # engine#2175 review: a script's short header sits BEHIND its `#!`. The first version of strip_short_header()
+    # looked at line 1 only, so the stub was not found, the block went in after the shebang, and the file ended up
+    # with two notices -- the same defect one file type over.
+    with open(os.path.join(repo, "short.sh"), "w") as f:
+        f.write("#!/usr/bin/env bash\n" + NOTICE.splitlines()[0] + "\n" + NOTICE.splitlines()[1]
+                + "\n# short.sh -- a script\necho hi\n")
+    git(repo, "add", "short.sh")
+    rc = run(repo, "--fix")
+    check(rc.returncode == 0, "--fix repairs a short header behind a shebang", rc.stdout + rc.stderr)
+    got = open(os.path.join(repo, "short.sh")).read()
+    check(got.startswith("#!/usr/bin/env bash\n"),
+          "the shebang stays on line 1", got.splitlines()[0])
+    check(got.count("SPDX-License-Identifier") == 1,
+          "and exactly one SPDX line in the script too", f"{got.count('SPDX-License-Identifier')} found")
+    check("short.sh -- a script" in got, "and the script's own comment survives", got)
+    rc = run(repo, "--check")
+    check(rc.returncode == 0, "and --check is green for the script", rc.stdout + rc.stderr)
+    os.unlink(os.path.join(repo, "short.sh"))
+    git(repo, "add", "-A", "short.sh")
+    os.unlink(os.path.join(repo, "short.h"))
+    git(repo, "add", "-A", "short.h")
+
     # an ignored file is not source: it must not need a notice. The .gitignore itself is source (the
     # repo's own carries the notice and is tracked), so it gets one here too.
     with open(os.path.join(repo, ".gitignore"), "w") as f:

tools/copyright.py

@@ -167,6 +167,21 @@ def has_notice(text):
     return NOTICE[0] in head and NOTICE[1] in head and "Licensed under the Apache License, Version 2.0" in head
 
 
+def short_notice(text):
+    """True when the header carries Copyright + SPDX but not the licence block.
+
+    engine#2175: `notice_block()` returns the 16-line form for a known style, so a header that keeps only
+    `// Copyright ...` and `// SPDX-License-Identifier: Apache-2.0` and then jumps to the file's own
+    description reads as MISSING -- and the message says "missing copyright notice", which sends the reader
+    looking for a notice that is on line 1. Measured on engine#2175 / #2173 (their audio headers) against
+    #2133 (the same work, full block): copyright present, SPDX present, "Licensed under the Apache License"
+    absent. Naming the shape is the difference between a 15-line paste and a hunt.
+    """
+    head = "\n".join(text.splitlines()[:NOTICE_WINDOW])
+    return (NOTICE[0] in head and NOTICE[1] in head
+            and "Licensed under the Apache License, Version 2.0" not in head)
+
+
 def duplicated_notice(text):
     """True when the Copyright + SPDX pair appears twice inside the header window.
 
@@ -367,7 +382,12 @@ def listing(*args):
     for path in unknown:
         print("no comment style known (add one to LINE_STYLES/BLOCK_STYLES or EXEMPT):", path)
     for path in missing:
-        print("missing copyright notice:", path)
+        # engine#2175: a short header is not a missing notice. Say which, so the fix is a paste rather than a
+        # search -- and keep the exit code the same (this is still a failure: the licence block is required).
+        if short_notice(open(path, encoding="utf-8", errors="replace").read()):
+            print("copyright + SPDX present but the licence block is not (paste the 16-line form):", path)
+        else:
+            print("missing copyright notice:", path)
     if missing or unknown or duplicated:
         print("run: python3 tools/copyright.py --fix", file=sys.stderr)
         return 1