refactor(git-gate): replace regex frontmatter edit with parse→mutate→serialize

The regex-based _insert_host_key_in_frontmatter had a bug: child_indent
was overwritten on every line with indent > repo_indent (including
grandchildren), so host_key landed inside the key: block instead of at
the repo level.

Replace with a clean parse → mutate → re-serialize approach:
- Add serialize_yaml_subset() to yaml_subset.py (block-style, 2-space
  indent, round-trips through parse_yaml_subset)
- Replace _insert_host_key_in_frontmatter + _update_frontmatter_in_file
  with _add_host_key_to_frontmatter that parses the frontmatter dict,
  sets fm["git-gate"]["repos"][name]["host_key"], and re-serializes
- _find_and_update_bottle_file simplified to read file → delegate to
  _add_host_key_to_frontmatter → write if changed
This commit is contained in:
2026-07-09 17:21:57 +00:00
parent 61740cdb6a
commit d496e30681
4 changed files with 342 additions and 187 deletions
+70 -101
View File
@@ -8,10 +8,9 @@ from pathlib import Path
from unittest.mock import MagicMock, patch
from bot_bottle.git_gate_host_key import (
_add_host_key_to_frontmatter,
_find_and_update_bottle_file,
_insert_host_key_in_frontmatter,
_prompt_tty,
_update_frontmatter_in_file,
fetch_host_key,
preflight_host_keys,
)
@@ -145,96 +144,11 @@ class TestPromptTty(unittest.TestCase):
# ---------------------------------------------------------------------------
# _insert_host_key_in_frontmatter
# _add_host_key_to_frontmatter
# ---------------------------------------------------------------------------
class TestInsertHostKeyInFrontmatter(unittest.TestCase):
_FM = """\
git-gate:
repos:
myrepo:
url: ssh://git@host/org/repo.git
key:
provider: static
path: /dev/null
"""
def test_inserts_host_key_after_last_child(self) -> None:
result = _insert_host_key_in_frontmatter(self._FM, "myrepo", "ssh-ed25519 AAAA")
self.assertIn(" host_key: ssh-ed25519 AAAA", result)
def test_no_change_when_repo_not_found(self) -> None:
result = _insert_host_key_in_frontmatter(self._FM, "other", "ssh-ed25519 AAAA")
self.assertEqual(self._FM, result)
def test_no_change_when_host_key_already_present(self) -> None:
fm = self._FM + " host_key: ssh-ed25519 EXISTING\n"
result = _insert_host_key_in_frontmatter(fm, "myrepo", "ssh-ed25519 NEW")
self.assertNotIn("NEW", result)
self.assertEqual(fm, result)
def test_preserves_other_repos(self) -> None:
fm = """\
git-gate:
repos:
first:
url: ssh://git@host/org/first.git
key:
provider: static
path: /dev/null
second:
url: ssh://git@host/org/second.git
key:
provider: static
path: /dev/null
"""
result = _insert_host_key_in_frontmatter(fm, "first", "ssh-ed25519 AAAA")
self.assertIn("host_key: ssh-ed25519 AAAA", result)
self.assertIn("second:", result)
def test_detects_child_indent_from_existing_keys(self) -> None:
# 4-space indent (non-standard but valid)
fm = """\
git-gate:
repos:
myrepo:
url: ssh://git@host/org/repo.git
key:
provider: static
path: /dev/null
"""
result = _insert_host_key_in_frontmatter(fm, "myrepo", "ssh-ed25519 AAAA")
self.assertIn(" host_key: ssh-ed25519 AAAA", result)
def test_blank_lines_inside_block_do_not_end_scan(self) -> None:
fm = """\
git-gate:
repos:
myrepo:
url: ssh://git@host/org/repo.git
key:
provider: static
path: /dev/null
"""
result = _insert_host_key_in_frontmatter(fm, "myrepo", "ssh-ed25519 AAAA")
self.assertIn("host_key: ssh-ed25519 AAAA", result)
def test_repo_with_no_children_uses_default_indent(self) -> None:
# Edge case: repo entry has no child keys at all (empty block).
fm = "git-gate:\n repos:\n myrepo:\n"
result = _insert_host_key_in_frontmatter(fm, "myrepo", "ssh-ed25519 AAAA")
# Should fall back to repo_indent+2 = 4
self.assertIn(" host_key: ssh-ed25519 AAAA", result)
# ---------------------------------------------------------------------------
# _update_frontmatter_in_file
# ---------------------------------------------------------------------------
class TestUpdateFrontmatterInFile(unittest.TestCase):
class TestAddHostKeyToFrontmatter(unittest.TestCase):
_FILE = """\
---
git-gate:
@@ -248,23 +162,78 @@ git-gate:
# Body text here
"""
def test_updates_frontmatter_preserves_body(self) -> None:
result = _update_frontmatter_in_file(self._FILE, "myrepo", "ssh-ed25519 AAAA")
def test_inserts_host_key_at_repo_level(self) -> None:
result = _add_host_key_to_frontmatter(self._FILE, "myrepo", "ssh-ed25519 AAAA")
self.assertIn("host_key: ssh-ed25519 AAAA", result)
# Must be a sibling of `url` and `key`, not nested inside `key`
lines = result.splitlines()
hk_line = next(l for l in lines if "host_key" in l)
url_line = next(l for l in lines if "url:" in l)
self.assertEqual(
len(hk_line) - len(hk_line.lstrip()),
len(url_line) - len(url_line.lstrip()),
)
def test_preserves_body(self) -> None:
result = _add_host_key_to_frontmatter(self._FILE, "myrepo", "ssh-ed25519 AAAA")
self.assertIn("# Body text here", result)
def test_preserves_other_repos(self) -> None:
text = """\
---
git-gate:
repos:
first:
url: ssh://git@host/org/first.git
key:
provider: static
path: /dev/null
second:
url: ssh://git@host/org/second.git
key:
provider: static
path: /dev/null
---
"""
result = _add_host_key_to_frontmatter(text, "first", "ssh-ed25519 AAAA")
self.assertIn("host_key: ssh-ed25519 AAAA", result)
self.assertIn("second:", result)
def test_no_change_when_host_key_already_present(self) -> None:
text = self._FILE.replace(
" path: /dev/null\n",
" path: /dev/null\n host_key: ssh-ed25519 EXISTING\n",
)
result = _add_host_key_to_frontmatter(text, "myrepo", "ssh-ed25519 NEW")
self.assertNotIn("NEW", result)
def test_no_change_when_repo_not_found(self) -> None:
result = _add_host_key_to_frontmatter(self._FILE, "other", "ssh-ed25519 AAAA")
self.assertEqual(self._FILE, result)
def test_no_change_without_frontmatter(self) -> None:
text = "No frontmatter here\n"
result = _update_frontmatter_in_file(text, "myrepo", "ssh-ed25519 AAAA")
result = _add_host_key_to_frontmatter(text, "myrepo", "ssh-ed25519 AAAA")
self.assertEqual(text, result)
def test_no_change_when_repo_not_in_file(self) -> None:
result = _update_frontmatter_in_file(self._FILE, "other", "ssh-ed25519 AAAA")
self.assertEqual(self._FILE, result)
def test_no_change_when_no_closing_delimiter(self) -> None:
text = "---\ngit-gate:\n repos:\n myrepo:\n url: x\n"
result = _update_frontmatter_in_file(text, "myrepo", "ssh-ed25519 AAAA")
result = _add_host_key_to_frontmatter(text, "myrepo", "ssh-ed25519 AAAA")
self.assertEqual(text, result)
def test_no_change_when_git_gate_missing(self) -> None:
text = "---\nenv:\n FOO: bar\n---\n"
result = _add_host_key_to_frontmatter(text, "myrepo", "ssh-ed25519 AAAA")
self.assertEqual(text, result)
def test_no_change_when_repos_missing(self) -> None:
text = "---\ngit-gate:\n foo: bar\n---\n"
result = _add_host_key_to_frontmatter(text, "myrepo", "ssh-ed25519 AAAA")
self.assertEqual(text, result)
def test_no_change_when_repo_entry_not_a_dict(self) -> None:
text = "---\ngit-gate:\n repos:\n - item\n---\n"
result = _add_host_key_to_frontmatter(text, "myrepo", "ssh-ed25519 AAAA")
self.assertEqual(text, result)
@@ -372,9 +341,9 @@ git-gate:
os.chmod(path, 0o644)
def test_skips_when_update_produces_no_change(self) -> None:
# _update_frontmatter_in_file returns the original text unchanged
# (e.g. because _insert_host_key_in_frontmatter can't locate the
# entry in the raw text). Line 174 must `continue` rather than write.
# _add_host_key_to_frontmatter returns the original text unchanged
# (e.g. because the entry is absent from the parsed structure).
# _find_and_update_bottle_file must continue rather than write.
content = """\
---
git-gate:
@@ -389,7 +358,7 @@ git-gate:
with tempfile.TemporaryDirectory() as d:
path = self._write_bottle(d, "dev", content)
with patch(
"bot_bottle.git_gate_host_key._update_frontmatter_in_file",
"bot_bottle.git_gate_host_key._add_host_key_to_frontmatter",
return_value=content, # no change
):
result = _find_and_update_bottle_file(Path(d), "myrepo", "ssh-ed25519 AAAA")
+137 -1
View File
@@ -6,7 +6,7 @@ import textwrap
import unittest
from bot_bottle.yaml_subset import YamlSubsetError
from bot_bottle.yaml_subset import parse_frontmatter, parse_yaml_subset
from bot_bottle.yaml_subset import parse_frontmatter, parse_yaml_subset, serialize_yaml_subset
def _y(s: str):
@@ -457,5 +457,141 @@ class TestEdgeAndErrorBranches(unittest.TestCase):
self.assertEqual(({}, ""), parse_frontmatter(""))
# ---------------------------------------------------------------------------
# serialize_yaml_subset
# ---------------------------------------------------------------------------
class TestSerializeYamlSubset(unittest.TestCase):
def _roundtrip(self, text: str) -> None:
"""Parse then serialize then re-parse; result must equal original parse."""
parsed = parse_yaml_subset(text)
serialized = serialize_yaml_subset(parsed)
self.assertEqual(parsed, parse_yaml_subset(serialized))
# --- scalar values ---------------------------------------------------------
def test_string_bare(self) -> None:
self.assertEqual("key: value\n", serialize_yaml_subset({"key": "value"}))
def test_string_empty_is_quoted(self) -> None:
out = serialize_yaml_subset({"key": ""})
self.assertIn("key: ''", out)
self._roundtrip("key: ''\n")
def test_string_true_like_is_quoted(self) -> None:
out = serialize_yaml_subset({"key": "true"})
self.assertIn("'true'", out)
self.assertEqual({"key": "true"}, parse_yaml_subset(out))
def test_string_int_like_is_quoted(self) -> None:
out = serialize_yaml_subset({"key": "42"})
self.assertIn("'42'", out)
self.assertEqual({"key": "42"}, parse_yaml_subset(out))
def test_string_special_start_char_is_quoted(self) -> None:
for ch in ('"', "'", "[", "{", "!", "&", "*", "#", "|", ">"):
out = serialize_yaml_subset({"key": ch + "rest"})
self.assertIn("'", out, msg=f"expected quoting for {ch!r}")
def test_none_emits_null(self) -> None:
self.assertEqual("key: null\n", serialize_yaml_subset({"key": None}))
def test_bool_true(self) -> None:
self.assertEqual("key: true\n", serialize_yaml_subset({"key": True}))
def test_bool_false(self) -> None:
self.assertEqual("key: false\n", serialize_yaml_subset({"key": False}))
def test_int(self) -> None:
self.assertEqual("key: 42\n", serialize_yaml_subset({"key": 42}))
# --- nested dict -----------------------------------------------------------
def test_nested_dict(self) -> None:
data: dict[str, object] = {"outer": {"inner": "val"}}
out = serialize_yaml_subset(data)
self.assertIn("outer:", out)
self.assertIn(" inner: val", out)
self.assertEqual(data, parse_yaml_subset(out))
def test_empty_dict_value_inline(self) -> None:
out = serialize_yaml_subset({"key": {}})
self.assertIn("key: {}", out)
# --- lists -----------------------------------------------------------------
def test_list_of_scalars(self) -> None:
data: dict[str, object] = {"items": ["a", "b", "c"]}
out = serialize_yaml_subset(data)
self.assertEqual(data, parse_yaml_subset(out))
def test_empty_list_value_inline(self) -> None:
out = serialize_yaml_subset({"key": []})
self.assertIn("key: []", out)
def test_list_of_mappings(self) -> None:
data: dict[str, object] = {"items": [{"name": "alpha", "val": 1}, {"name": "beta", "val": 2}]}
out = serialize_yaml_subset(data)
self.assertEqual(data, parse_yaml_subset(out))
def test_list_item_mapping_with_nested_dict(self) -> None:
data: dict[str, object] = {"items": [{"name": "x", "sub": {"k": "v"}}]}
out = serialize_yaml_subset(data)
self.assertEqual(data, parse_yaml_subset(out))
def test_list_item_first_key_is_nested_dict(self) -> None:
# Covers the branch where the FIRST key of a mapping list item
# has a nested dict/list value (not a scalar).
data: dict[str, object] = {"items": [{"cfg": {"a": 1, "b": 2}, "name": "x"}]}
out = serialize_yaml_subset(data)
self.assertEqual(data, parse_yaml_subset(out))
def test_list_item_mapping_continuation_keys(self) -> None:
data: dict[str, object] = {"items": [{"a": 1, "b": 2, "c": 3}]}
out = serialize_yaml_subset(data)
self.assertEqual(data, parse_yaml_subset(out))
# --- empty / trivial -------------------------------------------------------
def test_empty_dict(self) -> None:
self.assertEqual("", serialize_yaml_subset({}))
# --- realistic bottle frontmatter round-trip --------------------------------
def test_bottle_frontmatter_round_trip(self) -> None:
text = textwrap.dedent("""\
git-gate:
repos:
myrepo:
url: ssh://git@host/org/repo.git
key:
provider: static
path: /dev/null
""")
self._roundtrip(text)
def test_host_key_inserted_at_repo_level(self) -> None:
"""The canonical use-case: host_key lands as a sibling of url/key."""
from typing import cast
text = textwrap.dedent("""\
git-gate:
repos:
myrepo:
url: ssh://git@host/org/repo.git
key:
provider: static
path: /dev/null
""")
data = parse_yaml_subset(text)
repos = cast(dict[str, object], cast(dict[str, object], data["git-gate"])["repos"])
cast(dict[str, object], repos["myrepo"])["host_key"] = "ssh-ed25519 AAAA"
out = serialize_yaml_subset(data)
parsed_back = parse_yaml_subset(out)
repo = cast(dict[str, object], cast(dict[str, object], cast(dict[str, object], parsed_back["git-gate"])["repos"])["myrepo"])
self.assertEqual("ssh-ed25519 AAAA", repo["host_key"])
self.assertIn("provider", cast(dict[str, object], repo["key"]))
if __name__ == "__main__":
unittest.main()