From bd639314967d964f1116ddb757986d0b00e79328 Mon Sep 17 00:00:00 2001
From: Matěj Cepl <mcepl@cepl.eu>
Date: Wed, 20 Jul 2022 18:10:52 +0200
Subject: Generate actual content of commits.

Should work even with binary files.
---
 osc_fast_export.py | 56 ++++++++++++++++++++++++++++++++++++++++++++----------
 1 file changed, 46 insertions(+), 10 deletions(-)

diff --git a/osc_fast_export.py b/osc_fast_export.py
index 2bfa4d6..0042100 100755
--- a/osc_fast_export.py
+++ b/osc_fast_export.py
@@ -4,12 +4,18 @@
 import collections
 import configparser
 from datetime import datetime
+import logging
+import os.path
 import pathlib
 import subprocess
+import sys
 from typing import 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")
+
 authorsfile = pathlib.Path(".git", "authorsfile.txt")
 
 # For reading section-less config files
@@ -41,8 +47,9 @@ def osc_log() -> List[LogEntry]:
         log_str = subprocess.run(
             ["osc", "log", "--xml"], check=True, text=True, stdout=subprocess.PIPE
         ).stdout
-    except subprocess.CalledProcessError:
-        raise
+    except subprocess.CalledProcessError as exc:
+        raise RuntimeError(f"Cannot collect log of the package!") from exc
+
     tree = ET.fromstring(log_str)
     log_list = [
         LogEntry(
@@ -58,27 +65,56 @@ def osc_log() -> List[LogEntry]:
     return log_list
 
 
-def export_data(dt: str) -> str:
-    return f"data {len(dt)}\n{dt}"
+def checkout_revision(rev: int):
+    NULL = open(os.devnull, "wb")
+    try:
+        subprocess.check_call(
+            ["osc", "up", "-r", f"{rev}"], stdout=NULL, stderr=subprocess.PIPE
+        )
+        subprocess.check_call(["osc", "clean"], stdout=NULL, stderr=subprocess.PIPE)
+    except subprocess.CalledProcessError as exc:
+        raise RuntimeError(f"Cannot checkout revision {rev}!") from exc
 
 
 def print_export(entry: LogEntry) -> int:
     mark = entry.rev
     author = authors.get(entry.author, "<none>")
-    date = int(datetime.timestamp(entry.date))
 
+    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} {date} +0000")
-    print(export_data(entry.msg))
+    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}")
-    # The following is dummy, needs to be replaced by actual content of each commit.
+
+    # Create actual content of the commit
+    # It is easier just to wipe out everything and include files again.
     print("deleteall")
-    print("M 644 inline tralala")
-    print(export_data("tralalala"))
+    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
-- 
cgit