🚧 Update commit message format

This commit is contained in:
Jonas Bengtsson
2026-09-12 11:49:03 +02:00
parent 17596907f5
commit e773703ed3
8 changed files with 153 additions and 82 deletions
+2 -1
View File
@@ -18,6 +18,7 @@ See `functional specification.md` for the full specification.
- Security validation with forbidden patterns (abort by default) - Security validation with forbidden patterns (abort by default)
- Pluggable author resolvers with priority chain and confidence levels - Pluggable author resolvers with priority chain and confidence levels
- Per-file attribution recorded in the commit message - Per-file attribution recorded in the commit message
- One-line-per-file commit entries in `git diff --name-status` style (letter, path, author, last-modified time), with a configurable separator (`commit.file_separator`, default `;`)
- Structured commit trailers (`Tracker-Run`, `Tracker-Host`, ...) - Structured commit trailers (`Tracker-Run`, `Tracker-Host`, ...)
- Repository-level lock with stale-lock detection (stored outside the repo) - Repository-level lock with stale-lock detection (stored outside the repo)
- Rotating file logs plus console logging - Rotating file logs plus console logging
@@ -69,7 +70,7 @@ scope is defined by Section 44, *Definition of Done – Initial Release*.
- [x] Configurable author resolver chain with confidence levels and `UNKNOWN` fallback; sidecar metadata resolver implemented (FR-006, FR-007, Sections 9 – 11) - [x] Configurable author resolver chain with confidence levels and `UNKNOWN` fallback; sidecar metadata resolver implemented (FR-006, FR-007, Sections 9 – 11)
- [x] Per-file attribution in the commit message (FR-008, FR-012) - [x] Per-file attribution in the commit message (FR-008, FR-012)
- [x] Automatic commit generation (FR-009) - [x] Automatic commit generation (FR-009)
- [x] Structured commit message: summary counts, per-file details, `Tracker-*` trailers (FR-010, FR-011, Sections 15/16/36) - [x] Structured commit message: summary counts, one-line-per-file entries with change letter / path / author / local-time mtime, `Tracker-*` trailers (FR-010, FR-011, Sections 15/16/36)
- [x] No empty commits (FR-014) - [x] No empty commits (FR-014)
- [x] Rotating application logs, console + file (FR-015, Sections 22 – 23) - [x] Rotating application logs, console + file (FR-015, Sections 22 – 23)
- [x] Cross-platform code base: `pathlib`, argument arrays, UTF-8/NUL-safe Git parsing (FR-016, FR-029) - [x] Cross-platform code base: `pathlib`, argument arrays, UTF-8/NUL-safe Git parsing (FR-016, FR-029)
+6
View File
@@ -36,6 +36,12 @@ author_resolution:
minimum_confidence: probable minimum_confidence: probable
fallback: unknown fallback: unknown
commit:
# Separator between the parts of a one-line file entry in commit
# messages (Section 15). Must not occur inside paths, author
# identifiers or timestamps.
file_separator: ";"
logging: logging:
level: INFO level: INFO
console: true console: true
+31 -30
View File
@@ -526,10 +526,21 @@ If the staged change set does not match the expected change set, the commit shal
Commit messages shall be human-readable while retaining a predictable structure that can later be parsed automatically. Commit messages shall be human-readable while retaining a predictable structure that can later be parsed automatically.
Each changed file shall occupy exactly one line, in the style of `git diff --name-status`. A single-letter change type is followed by the file path, the last modifying user and the filesystem last-modified time, with the parts joined by a configurable separator (see Section 26, `commit.file_separator`; default `;`).
Line grammar:
```text
A <sep> <path> <sep> <author> <sep> <modified-at> added
M <sep> <path> <sep> <author> <sep> <modified-at> modified
R <sep> <old-path> -> <new-path> renamed (bare)
D <sep> <path> deleted (bare)
```
Recommended format: Recommended format:
```text ```text
SCADA automatic change capture - 2026-09-11 18:00:03 SCADA automatic change capture - 2026-09-11 18:00:03+02:00
Summary: Summary:
Added: 2 Added: 2
@@ -537,31 +548,19 @@ Summary:
Deleted: 1 Deleted: 1
Renamed: 0 Renamed: 0
Added: Changes:
Stations/PS101/pump2.xml A ; Stations/PS101/pump2.xml ; DOMAIN\jsmith ; 2026-09-11 15:41:22+02:00
Last modified by: DOMAIN\jsmith A ; Stations/PS104/settings.xml ; DOMAIN\adoe ; 2026-09-11 16:02:14+02:00
Modified at: 2026-09-11 15:41:22 M ; Stations/PS101/project.xml ; DOMAIN\jsmith ; 2026-09-11 15:42:09+02:00
M ; Stations/PS103/alarm.xml ; DOMAIN\adoe ; 2026-09-11 16:13:54+02:00
Stations/PS104/settings.xml M ; System/global.xml ; UNKNOWN ; 2026-09-11 16:20:01+02:00
Last modified by: DOMAIN\adoe D ; Stations/PS027/old-config.xml
Modified at: 2026-09-11 16:02:14
Modified:
Stations/PS101/project.xml
Last modified by: DOMAIN\jsmith
Modified at: 2026-09-11 15:42:09
Stations/PS103/alarm.xml
Last modified by: DOMAIN\adoe
Modified at: 2026-09-11 16:13:54
System/global.xml
Last modified by: UNKNOWN
Deleted:
Stations/PS027/old-config.xml
``` ```
Deleted files are listed without author or time information (FR-013). Renames use git's `old -> new` arrow notation and stay bare: a pure rename preserves the original modification time, so a timestamp would be misleading. The time column is the filesystem last-modified time (Section 12: `file_timestamp`) rendered in local time with an explicit UTC offset, matching the log timestamp style (Section 22).
The one-line format keeps the message proportional to the number of changed files, which matters for large collections.
--- ---
# 16. Commit Subject # 16. Commit Subject
@@ -609,12 +608,9 @@ One execution may contain changes made by multiple individuals.
Example: Example:
```text ```text
Modified: Changes:
plant1.xml M ; plant1.xml ; DOMAIN\jsmith ; 2026-09-11 15:42:09+02:00
Last modified by: DOMAIN\jsmith M ; plant2.xml ; DOMAIN\adoe ; 2026-09-11 16:02:14+02:00
plant2.xml
Last modified by: DOMAIN\adoe
``` ```
The application shall therefore store attribution per file rather than per commit. The application shall therefore store attribution per file rather than per commit.
@@ -838,6 +834,11 @@ author_resolution:
minimum_confidence: probable minimum_confidence: probable
fallback: unknown fallback: unknown
commit:
# Separator between the parts of a one-line file entry (Section 15).
# Must not occur inside paths, author identifiers or timestamps.
file_separator: ";"
logging: logging:
level: INFO level: INFO
console: true console: true
+6 -1
View File
@@ -171,7 +171,12 @@ class ApplicationController:
"Author resolved: %s -> %s", record.path, result.identifier "Author resolved: %s -> %s", record.path, result.identifier
) )
message = build_commit_message(kept, run, app_version=_APP_VERSION) message = build_commit_message(
kept,
run,
app_version=_APP_VERSION,
file_separator=self.config.commit.file_separator,
)
if dry_run: if dry_run:
log.info("Dry-run: repository left unmodified") log.info("Dry-run: repository left unmodified")
+39 -32
View File
@@ -2,13 +2,17 @@
Human-readable sections first, Git trailer block last so that Human-readable sections first, Git trailer block last so that
`git interpret-trailers` can extract the structured metadata. `git interpret-trailers` can extract the structured metadata.
Each changed file occupies exactly ONE line in `git diff --name-status`
style: a change letter followed by path, author and last-modified time,
joined by a configurable separator (commit.file_separator, default ';').
""" """
from __future__ import annotations from __future__ import annotations
import datetime as dt import datetime as dt
from git_auto_tracker.changes.models import ChangeSet, TrackerRun from git_auto_tracker.changes.models import ChangeRecord, ChangeSet, ChangeType, TrackerRun
_SUMMARY_LABELS = ( _SUMMARY_LABELS = (
("Added:", "added"), ("Added:", "added"),
@@ -17,6 +21,13 @@ _SUMMARY_LABELS = (
("Renamed:", "renamed"), ("Renamed:", "renamed"),
) )
_GIT_LETTER = {
ChangeType.ADDED: "A",
ChangeType.MODIFIED: "M",
ChangeType.DELETED: "D",
ChangeType.RENAMED: "R",
}
def _format_ts(moment: dt.datetime | None) -> str: def _format_ts(moment: dt.datetime | None) -> str:
"""Render an aware datetime in local time with an explicit offset. """Render an aware datetime in local time with an explicit offset.
@@ -30,20 +41,28 @@ def _format_ts(moment: dt.datetime | None) -> str:
return moment.astimezone().isoformat(sep=" ", timespec="seconds") return moment.astimezone().isoformat(sep=" ", timespec="seconds")
def _author_lines(record) -> list[str]: # noqa: ANN001 - ChangeRecord, avoid cycle def _file_line(record: ChangeRecord, separator: str) -> str:
lines: list[str] = [] """One line per file, git --name-status style (Section 15).
author = record.author
identifier = author.identifier if author is not None else "UNKNOWN" Grammar:
lines.append(f" Last modified by: {identifier}") A <sep> <path> <sep> <author> <sep> <mtime> added
# Filesystem last-modified time (Section 12: file_timestamp); rendered M <sep> <path> <sep> <author> <sep> <mtime> modified
# in local time like the log timestamps. Resolver-provided timestamps R <sep> <old-path> -> <new-path> renamed (bare)
# (author.timestamp) stay authoritative and are shown as well. D <sep> <path> deleted (bare)
file_ts = _format_ts(record.file_timestamp)
if file_ts: The filesystem last-modified time (Section 12: file_timestamp) is used
lines.append(f" Modified at: {file_ts}") for the time column, rendered in local time like the log timestamps.
if author is not None and author.timestamp is not None: Deleted and renamed entries stay bare: the file is gone (FR-013), and a
lines.append(f" Changed at: {_format_ts(author.timestamp)}") pure rename preserves the original mtime so a timestamp would mislead.
return lines """
letter = _GIT_LETTER[record.change_type]
if record.change_type is ChangeType.RENAMED:
return f"{letter} {separator} {record.previous_path or record.path} -> {record.path}"
if record.change_type is ChangeType.DELETED:
return f"{letter} {separator} {record.path}"
author = record.author.identifier if record.author is not None else "UNKNOWN"
parts = [letter, record.path, author, _format_ts(record.file_timestamp)]
return f" {separator} ".join(parts)
def build_commit_message( def build_commit_message(
@@ -51,6 +70,7 @@ def build_commit_message(
run: TrackerRun, run: TrackerRun,
collected_at: dt.datetime | None = None, collected_at: dt.datetime | None = None,
app_version: str = "1.0.0", app_version: str = "1.0.0",
file_separator: str = ";",
) -> str: ) -> str:
"""Build the complete commit message per Sections 15/16/36.""" """Build the complete commit message per Sections 15/16/36."""
counts = change_set.summary_counts() counts = change_set.summary_counts()
@@ -64,23 +84,10 @@ def build_commit_message(
lines.append(f" {label} {counts[key]}") lines.append(f" {label} {counts[key]}")
lines.append("") lines.append("")
for heading, records in ( lines.append("Changes:")
("Added:", change_set.added), for record in change_set.all_records():
("Modified:", change_set.modified), lines.append(f" {_file_line(record, file_separator)}")
("Renamed:", change_set.renamed), lines.append("")
("Deleted:", change_set.deleted),
):
if not records:
continue
lines.append(heading)
for record in records:
if record.change_type.value == "RENAMED" and record.previous_path:
lines.append(f" {record.previous_path} -> {record.path}")
else:
lines.append(f" {record.path}")
if record.needs_author:
lines.extend(_author_lines(record))
lines.append("")
# Structured trailers (Section 36) # Structured trailers (Section 36)
lines.append(f"Tracker-Version: {app_version}") lines.append(f"Tracker-Version: {app_version}")
+20
View File
@@ -90,6 +90,24 @@ class LoggingConfig:
backup_count: int = 10 backup_count: int = 10
@dataclass
class CommitConfig:
"""Commit-message rendering options (Sections 15 and 16)."""
#: Separator between the parts of a one-line file entry. Must not be
#: empty; must not occur inside paths, author identifiers or timestamps.
file_separator: str = ";"
@classmethod
def from_dict(cls, data: dict[str, Any]) -> CommitConfig:
separator = data.get("file_separator", ";")
if not isinstance(separator, str) or not separator.strip():
raise ConfigurationError("commit.file_separator must be a non-empty string")
if separator in ("\n", "\r"):
raise ConfigurationError("commit.file_separator must not be a line break")
return cls(file_separator=separator)
@dataclass @dataclass
class AppConfig: class AppConfig:
repository_path: Path repository_path: Path
@@ -101,6 +119,7 @@ class AppConfig:
author_resolution: AuthorResolutionConfig author_resolution: AuthorResolutionConfig
logging: LoggingConfig logging: LoggingConfig
timezone: str | None = None timezone: str | None = None
commit: CommitConfig = field(default_factory=CommitConfig)
@classmethod @classmethod
def from_dict(cls, data: dict[str, Any], config_dir: Path) -> AppConfig: def from_dict(cls, data: dict[str, Any], config_dir: Path) -> AppConfig:
@@ -152,6 +171,7 @@ class AppConfig:
max_size_mb=int(logging_cfg.get("max_size_mb", 20)), max_size_mb=int(logging_cfg.get("max_size_mb", 20)),
backup_count=int(logging_cfg.get("backup_count", 10)), backup_count=int(logging_cfg.get("backup_count", 10)),
), ),
commit=CommitConfig.from_dict(data.get("commit", {})),
timezone=app_cfg.get("timezone"), timezone=app_cfg.get("timezone"),
) )
except ConfigurationError: except ConfigurationError:
+19 -6
View File
@@ -100,23 +100,36 @@ class TestRunCycle:
) )
_run(tracker_config, tmp_path) _run(tracker_config, tmp_path)
body = _git_out("log", "-1", "--format=%b", cwd=git_repo) body = _git_out("log", "-1", "--format=%b", cwd=git_repo)
assert "Last modified by: DOMAIN\\jsmith" in body # One-line entry: letter ; path ; author ; local-time mtime
assert " A ; config.xml ; DOMAIN\\jsmith ; " in body
def test_file_mtime_appears_in_commit_local_time( def test_file_mtime_appears_in_commit_local_time(
self, git_repo: Path, tracker_config: dict, tmp_path: Path self, git_repo: Path, tracker_config: dict, tmp_path: Path
) -> None: ) -> None:
"""Section 12 file_timestamp rendered as 'Modified at' (local time).""" """Section 12 file_timestamp on the one-line entry (local time)."""
import re as re_mod import re as re_mod
(git_repo / "config.xml").write_text("<c/>", encoding="utf-8") (git_repo / "config.xml").write_text("<c/>", encoding="utf-8")
_run(tracker_config, tmp_path) _run(tracker_config, tmp_path)
body = _git_out("log", "-1", "--format=%b", cwd=git_repo) body = _git_out("log", "-1", "--format=%b", cwd=git_repo)
# Local-time format with explicit +/-HH:MM offset (Section 15) # A ; path ; author ; local-time mtime with +/-HH:MM offset
assert re_mod.search( assert re_mod.search(
r"Modified at: \d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}[+-]\d{2}:\d{2}\b", body r"^ A ; config\.xml ; UNKNOWN ; "
r"\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}[+-]\d{2}:\d{2}$",
body,
flags=re_mod.MULTILINE,
) )
# The timestamp block sits directly under the changed file entry
assert " config.xml\n Last modified by: UNKNOWN\n Modified at:" in body def test_custom_separator_in_real_commit(
self, git_repo: Path, tracker_config: dict, tmp_path: Path
) -> None:
"""commit.file_separator flows from configuration into the commit."""
tracker_config["commit"] = {"file_separator": "|"}
(git_repo / "config.xml").write_text("<c/>", encoding="utf-8")
_run(tracker_config, tmp_path)
body = _git_out("log", "-1", "--format=%b", cwd=git_repo)
assert " A | config.xml | UNKNOWN | " in body
assert "; config.xml" not in body
def test_excluded_files_not_committed( def test_excluded_files_not_committed(
self, git_repo: Path, tracker_config: dict, tmp_path: Path self, git_repo: Path, tracker_config: dict, tmp_path: Path
+30 -12
View File
@@ -102,6 +102,7 @@ class TestConfig:
assert config.repository_path == Path(tracker_config["repository"]["path"]) assert config.repository_path == Path(tracker_config["repository"]["path"])
assert config.branch == "main" assert config.branch == "main"
assert config.author_resolution.minimum_confidence == "probable" assert config.author_resolution.minimum_confidence == "probable"
assert config.commit.file_separator == ";" # default
def test_missing_file(self, tmp_path: Path) -> None: def test_missing_file(self, tmp_path: Path) -> None:
with pytest.raises(ConfigurationError): with pytest.raises(ConfigurationError):
@@ -175,16 +176,19 @@ class TestCommitMessage:
assert lines[0].startswith("SCADA automatic change capture - ") assert lines[0].startswith("SCADA automatic change capture - ")
assert "Summary:" in lines assert "Summary:" in lines
assert "Added: 1" in message assert "Added: 1" in message
assert "Last modified by: DOMAIN\\jsmith" in message assert "Deleted: 1" in message
assert "Last modified by: UNKNOWN" in message # One-line-per-file entries with the default ';' separator (Section 15)
# Filesystem mtime rendered as 'Modified at' in local time with offset assert " A ; Stations/PS101/pump2.xml ; UNKNOWN ; " in message
assert " Modified at: " in message assert " M ; Stations/PS101/project.xml ; DOMAIN\\jsmith ; " in message
assert " M ; System/global.xml ; UNKNOWN ; " in message
assert " D ; Stations/PS027/old.xml" in message
# Local-time timestamps with explicit +/-HH:MM offset
assert re.search( assert re.search(
r"Modified at: \d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}[+-]\d{2}:\d{2}\b", message r"; \d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}[+-]\d{2}:\d{2}\b", message
) )
# Resolver-provided timestamp kept as a separate 'Changed at' line # One 'Changes:' section; no multi-line author blocks remain
assert " Changed at: " in message assert "Last modified by:" not in message
assert "Stations/PS027/old.xml" in message assert "Changes:" in message
# Trailers last (Section 36) # Trailers last (Section 36)
assert "Tracker-Version: 1.0.0" in message assert "Tracker-Version: 1.0.0" in message
assert "Tracker-Run: 01J7K8HYV84ZKFH1J4T9R8DZQN" in message assert "Tracker-Run: 01J7K8HYV84ZKFH1J4T9R8DZQN" in message
@@ -199,9 +203,12 @@ class TestCommitMessage:
deleted=[ChangeRecord(path="gone.xml", change_type=ChangeType.DELETED)] deleted=[ChangeRecord(path="gone.xml", change_type=ChangeType.DELETED)]
) )
message = build_commit_message(change_set, self._run()) message = build_commit_message(change_set, self._run())
idx = message.index("gone.xml") # Deleted entries are bare: letter, separator, path - no author/time
following = message[idx:].splitlines() assert " D ; gone.xml" in message
assert not any("Last modified by" in line for line in following[1:3]) assert "UNKNOWN" not in message
for line in message.splitlines():
if "gone.xml" in line:
assert line.count(";") == 1
def test_rename_rendered_with_arrow(self) -> None: def test_rename_rendered_with_arrow(self) -> None:
change_set = ChangeSet( change_set = ChangeSet(
@@ -212,4 +219,15 @@ class TestCommitMessage:
] ]
) )
message = build_commit_message(change_set, self._run()) message = build_commit_message(change_set, self._run())
assert "old.xml -> new.xml" in message assert " R ; old.xml -> new.xml" in message
def test_custom_separator(self) -> None:
change_set = ChangeSet(
added=[ChangeRecord(path="pump.xml", change_type=ChangeType.ADDED)],
deleted=[ChangeRecord(path="gone.xml", change_type=ChangeType.DELETED)],
)
run = self._run()
message = build_commit_message(change_set, run, file_separator="|")
assert " A | pump.xml ; UNKNOWN ; " not in message
assert " A | pump.xml | UNKNOWN | " in message
assert " D | gone.xml" in message