#!/usr/bin/python3 # https://git-scm.com/book/en/v2/Git-and-Other-Systems-Migrating-to-Git#_custom_importer import collections import configparser from datetime import datetime import logging import os.path import pathlib from subprocess import CalledProcessError, PIPE, run import sys from typing import Dict, List import xml.etree.ElementTree as ET logging.basicConfig(format="%(levelname)s:%(funcName)s:%(message)s", level=logging.INFO) log = logging.getLogger("osc_fast_export") NULL = open(os.devnull, "wb") # For reading section-less config files # https://stackoverflow.com/a/2819788/164233 def FakeSecHead(fp): yield "[asection]\n" yield from fp class LogEntry( collections.namedtuple("LogEntry", ["rev", "md5", "author", "date", "msg"]) ): def __str__(self): return ( f"{self.rev}, {self.md5[:12]}, {authors.get(self.author, '')}," + f" {datetime.isoformat(self.date)}:\n{self.msg}" ) def get_authors() -> Dict[str, str]: config = configparser.ConfigParser() authors = {} authorsfile = pathlib.Path(".osc", "authorsfile.txt") if authorsfile.exists(): config.read_file(FakeSecHead(open(authorsfile))) authors = dict(config.items("asection")) return authors def osc_log() -> List[LogEntry]: try: log_pid = run(["osc", "log", "--xml"], check=True, text=True, stdout=PIPE, stderr=PIPE ) except CalledProcessError as exc: raise RuntimeError( f"Cannot collect log of the package!\nReported issue was:\n{log_pid.stderr}") from exc log_str = log_pid.stdout tree = ET.fromstring(log_str) log_list = [ LogEntry( int(entry.attrib["revision"]), entry.attrib["srcmd5"], entry.findtext("author"), datetime.strptime(entry.findtext("date"), "%Y-%m-%d %H:%M:%S"), entry.findtext("msg"), ) for entry in tree.iter("logentry") ] log_list.reverse() return log_list def checkout_revision(rev: int): try: osc_pid = run( ["osc", "up", "-r", f"{rev}"], stdout=PIPE, stderr=PIPE, check=True ) except CalledProcessError as exc: raise RuntimeError( f"Cannot checkout revision {rev}!\nReported issue was:\n{osc_pid.stderr}") from exc try: osc_pid = run(["osc", "clean"], stdout=PIPE, stderr=PIPE, check=True) except CalledProcessError as exc: raise RuntimeError( f"Cannot clean checkout!\nReported issue was:\n{osc_pid.stderr}") from exc def print_export(entry: LogEntry, authors: Dict[str, str]) -> int: mark = entry.rev author = authors.get(entry.author, "") checkout_revision(entry.rev) print("commit refs/heads/master") print(f"mark :{mark}") if entry.md5: print(f"original-oid {entry.md5}") print(f"committer {author} {entry.date:%s} +0000") print(f"data {len(entry.msg)}\n{entry.msg}") if last_mark: print(f"from :{last_mark}") # Create actual content of the commit # It is easier just to wipe out everything and include files again. print("deleteall") for dirpath, dirnames, filenames in os.walk("."): # TODO are osc notes (aka osc comment) a thing? dirpath = os.path.relpath(dirpath, ".") if dirpath.startswith(".osc"): continue # It seems git-fast-export doesn't export directories at all # and git-fast-import just creates them when needed. # if dirpath != '.': # # create directory # print(f'M 040000 inline {dirpath}') for fn in filenames: fname = os.path.relpath(os.path.join(dirpath, fn), ".") log.debug("dirpath = %s, fname = %s", dirpath, fname) fstat = f"{os.stat(fname).st_mode:o}" print(f"M {fstat} inline {fname}") with open(fname, "rb") as inf: dt = inf.read() print(f"data {len(dt)}") sys.stdout.flush() with os.fdopen(sys.stdout.fileno(), "wb", closefd=False) as stdout: stdout.write(dt) stdout.flush() print("") return mark if __name__ == "__main__": last_mark = None authors = get_authors() for logentry in osc_log(): last_mark = print_export(logentry, authors) print("done")