blame function is running slowly, and the results do not match the git blame command

未关闭
#1,401 0 条评论 0 个 reaction 已指派 0 人 在 GitHub 查看

还没有人认领这个 Issue。

评估

难度
4/5
预计耗时
3-5 天
新手友好度
35/100
Issue 类型
缺陷
描述清晰度
需要澄清
活跃度
停滞
技术栈
git, python
领域
devtools

调研方向

从提供的 Python 示例和 repo.blame 调用开始,将其输出和执行时间与所示的 git checkout 和 git blame 命令进行比较。调查 blame 结果为何不同,以及 pygit2 运行时间为何更长;当行为与 git blame 一致且执行时间得到改善时,即视为完成。

由索引模型根据 Issue 内容生成。

描述

I want to use pygit2 to trace back the vulnerability introduction time based on the fix patches of open-source components.

On Ubuntu, testing the Linux repository takes about 13 seconds.

time (git checkout 2734d6c1b1a089fb593ef6a23d4b70903526fe0c && git blame -L 3883,3886 kernel/trace/ring_buffer.c)
Updating files: 100% (77303/77303), done.
Note: switching to '2734d6c1b1a089fb593ef6a23d4b70903526fe0c'.

You are in 'detached HEAD' state. You can look around, make experimental
changes and commit them, and you can discard any commits you make in this
state without impacting any branches by switching back to a branch.

If you want to create a new branch to retain commits you create, you may
do so (now or later) by using -c with the switch command. Example:

  git switch -c <new-branch-name>

Or undo this operation with:

  git switch -

Turn off this advice by setting config variable advice.detachedHead to false

HEAD is now at 2734d6c1b1a0 Linux 5.14-rc2
bf41a158cacba (Steven Rostedt 2008-10-04 02:00:59 -0400 3883)   return reader->read == rb_page_commit(reader) &&
bf41a158cacba (Steven Rostedt 2008-10-04 02:00:59 -0400 3884)           (commit == reader ||
bf41a158cacba (Steven Rostedt 2008-10-04 02:00:59 -0400 3885)            (commit == head &&
bf41a158cacba (Steven Rostedt 2008-10-04 02:00:59 -0400 3886)             head->read == rb_page_commit(commit)));

real    0m13.343s
user    0m6.960s
sys     0m5.266s

However, the execution time of pygit2 is significantly slower than direct command-line operations, and the running results are inconsistent.

Here is my python code:

def show_commit_line_blame(repo_path, commit_hash, file_path, min_line_number, max_line_number):
    repo = pygit2.Repository(repo_path)
    blame = repo.blame(
        file_path, 
        newest_commit=commit_hash, 
        flags=pygit2.GIT_BLAME_FIRST_PARENT, 
        min_line=min_line_number, 
        max_line=max_line_number
    )
    
    for line_number in range(min_line_number, max_line_number+1):
        hunk = blame.for_line(line_number)
        blamed_commit = repo.get(hunk.final_commit_id)
        utc_time = datetime.fromtimestamp(blamed_commit.commit_time, tz=timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
        print(f"Line {line_number} in {file_path} last modified by commit {blamed_commit.id} on {utc_time}")

output:
Line 3883 in kernel/trace/ring_buffer.c last modified by commit 92b29b86fe2e183d44eb467e5e74a5f718ef2e43 on 2008-10-20T20:35:07Z
Line 3884 in kernel/trace/ring_buffer.c last modified by commit 92b29b86fe2e183d44eb467e5e74a5f718ef2e43 on 2008-10-20T20:35:07Z
Line 3885 in kernel/trace/ring_buffer.c last modified by commit 92b29b86fe2e183d44eb467e5e74a5f718ef2e43 on 2008-10-20T20:35:07Z
Line 3886 in kernel/trace/ring_buffer.c last modified by commit 92b29b86fe2e183d44eb467e5e74a5f718ef2e43 on 2008-10-20T20:35:07Z
used : 50.29366707801819s

environment:
ubuntu 22.04 (6.8.0-58-generic #60~22.04.1-Ubuntu SMP)
Python 3.10.12
pygit2 1.18.1

Could you give me some suggestions to improve execution efficiency?

主要语言
Python
星标
1.7k
派生
408
平均合并
2 天 57 分钟
30 天内合并 PR
7

贡献指南

打开贡献指南

从这里开始

  1. 先读完整个 Issue,再读项目的贡献指南。
  2. 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
  3. Fork 仓库,在一个分支上完成修改。
  4. 提交 Pull Request,并在描述里引用这个 Issue 编号。

libgit2/pygit2 的其他 Issue

查看 libgit2/pygit2 的全部 Issue

相似的 Issue

更多 Python Issue

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。