sync: SHANGHAI-U9 2026-08-19 tools-sync 增 504 恢复工具 recover_commits_via_gitea_api.py 并写进阶段1处置

2026-08-19 15:16:56 +08:00
parent 28307b12f6
commit 307bc38d36
2 changed files with 156 additions and 0 deletions
@@ -0,0 +1,147 @@
+"""Rebuild commits the git transport cannot fetch, using the Gitea HTTP API.
+
+Why this exists: git.rpchannel.com has repeatedly served `info/refs` and the
+whole HTTP API fine while `git-upload-pack` / `git-receive-pack` 504 (the
+server-side git process hangs -- a 32KB push body 504s just as a full fetch
+does). That leaves the local checkout unable to rebase onto a remote head it
+can see but cannot download.
+
+The API exposes the same objects the pack protocol would have shipped, so the
+missing commits can be rebuilt locally:
+
+    /api/v1/repos/<o>/<r>/git/commits/<sha>            author, committer, message, parents
+    /api/v1/repos/<o>/<r>/git/trees/<sha>?recursive=1  every path, mode and blob sha
+    /api/v1/repos/<o>/<r>/git/blobs/<sha>              base64 content
+
+This is a reconstruction, not a copy, so it is only trustworthy because it is
+self-verifying: every blob is accepted only if its SHA-1 matches what the
+server reported, and a commit is written only if the assembled object hashes to
+the exact sha being recovered. A mismatch aborts instead of writing something
+plausible.
+
+Two traps, both of which produce a silently wrong result rather than an error:
+  * `git update-index --index-info` input must be bytes. With text=True on
+    Windows the newlines become CRLF, every line is rejected, and write-tree
+    happily returns the EMPTY tree (4b825dc6...).
+  * Gitea fills the commit's `tree.sha` field with the COMMIT sha. Do not use it
+    as a check -- verify the final commit sha instead.
+
+Usage:
+    python recover_commits_via_gitea_api.py <repo_dir> <sha> [<sha> ...]
+
+Recover in parent-first order. Afterwards the shas are real local objects:
+    git update-ref refs/remotes/origin/<branch> <newest_sha>
+    git rebase <newest_sha>
+"""
+
+import base64
+import datetime
+import hashlib
+import json
+import os
+import re
+import subprocess
+import sys
+import urllib.request
+
+
+def api_base(repo_dir):
+    """Derive the Gitea API root from origin's URL."""
+    url = subprocess.run(("git", "-C", repo_dir, "remote", "get-url", "origin"),
+                         capture_output=True, text=True, check=True).stdout.strip()
+    m = re.match(r"https?://([^/]+)/([^/]+)/([^/]+?)(?:\.git)?/?$", url)
+    if not m:
+        sys.exit(f"cannot parse a Gitea owner/repo out of origin url: {url}")
+    host, owner, repo = m.groups()
+    return f"http://{host}/api/v1/repos/{owner}/{repo}"
+
+
+def get(url):
+    with urllib.request.urlopen(url, timeout=90) as f:
+        return json.load(f)
+
+
+def git_date(iso):
+    """'2026-07-30T22:19:56+08:00' -> ('1785...', '+0800')"""
+    dt = datetime.datetime.fromisoformat(iso)
+    total = int(dt.utcoffset().total_seconds())
+    sign = "+" if total >= 0 else "-"
+    total = abs(total)
+    return str(int(dt.timestamp())), f"{sign}{total // 3600:02d}{(total % 3600) // 60:02d}"
+
+
+def recover(sha, base, env, index_file):
+    meta = get(f"{base}/git/commits/{sha}")
+    tree = get(f"{base}/git/trees/{sha}?recursive=true&per_page=1000")
+    if tree.get("truncated"):
+        sys.exit(f"{sha[:8]}: tree listing truncated -- raise per_page")
+
+    blobs = [e for e in tree["tree"] if e["type"] == "blob"]
+    fetched = 0
+    for e in blobs:
+        present = subprocess.run(("git", "cat-file", "-e", e["sha"] + "^{object}"),
+                                 env=env, capture_output=True).returncode == 0
+        if present:
+            continue
+        raw = base64.b64decode(get(f"{base}/git/blobs/{e['sha']}")["content"])
+        got = hashlib.sha1(b"blob %d\0" % len(raw) + raw).hexdigest()
+        if got != e["sha"]:
+            sys.exit(f"blob mismatch for {e['path']}: {got} != {e['sha']}")
+        w = subprocess.run(("git", "hash-object", "-t", "blob", "-w", "--stdin"),
+                           env=env, input=raw, capture_output=True)
+        if w.stdout.decode().strip() != e["sha"]:
+            sys.exit(f"write-back drift for {e['path']}")
+        fetched += 1
+
+    if os.path.exists(index_file):
+        os.remove(index_file)
+    payload = "".join(f"{e['mode']} {e['sha']}\t{e['path']}\n" for e in blobs)
+    r = subprocess.run(("git", "update-index", "--index-info"), env=env,
+                       input=payload.encode("utf-8"), capture_output=True)
+    if r.returncode:
+        sys.exit(f"update-index failed: {r.stderr.decode()}")
+    tree_sha = subprocess.run(("git", "write-tree"), env=env,
+                              capture_output=True, text=True, check=True).stdout.strip()
+
+    c = meta["commit"]
+    ats, atz = git_date(c["author"]["date"])
+    cts, ctz = git_date(c["committer"]["date"])
+    head = f"tree {tree_sha}\n"
+    for p in meta.get("parents", []):
+        head += f"parent {p['sha']}\n"
+    head += (f"author {c['author']['name']} <{c['author']['email']}> {ats} {atz}\n"
+             f"committer {c['committer']['name']} <{c['committer']['email']}> {cts} {ctz}\n\n")
+
+    msg = c["message"]
+    for variant in (msg, msg + "\n", msg.rstrip("\n") + "\n", msg.rstrip("\n")):
+        body = (head + variant).encode("utf-8")
+        if hashlib.sha1(b"commit %d\0" % len(body) + body).hexdigest() == sha:
+            w = subprocess.run(("git", "hash-object", "-t", "commit", "-w", "--stdin"),
+                               env=env, input=body, capture_output=True)
+            if w.stdout.decode().strip() != sha:
+                sys.exit(f"{sha[:8]}: commit write drift")
+            print(f"OK   {sha[:8]}  tree={tree_sha[:8]}  blobs_fetched={fetched}")
+            return True
+    print(f"FAIL {sha[:8]}  no reconstruction hashes to this sha "
+          f"(signed commit, or the API is reporting different metadata)")
+    return False
+
+
+def main():
+    if len(sys.argv) < 3:
+        sys.exit(__doc__)
+    repo_dir, shas = sys.argv[1], sys.argv[2:]
+    git_dir = subprocess.run(("git", "-C", repo_dir, "rev-parse", "--absolute-git-dir"),
+                             capture_output=True, text=True, check=True).stdout.strip()
+    index_file = os.path.join(os.environ.get("TEMP", "."), "recover_gitea_index")
+    # A private index keeps the caller's working tree and real index untouched.
+    env = dict(os.environ, GIT_DIR=git_dir, GIT_INDEX_FILE=index_file)
+    base = api_base(repo_dir)
+    ok = all(recover(s, base, env, index_file) for s in shas)
+    if os.path.exists(index_file):
+        os.remove(index_file)
+    return 0 if ok else 1
+
+
+if __name__ == "__main__":
+    sys.exit(main())