“死了” Mercurial 中的目录
在我的 Hg 存储库（位于 Google Code）中，我有一个不应该存在的目录。 基本上我决定将所有内容移动到一个子目录中（我的所有内容都在根目录中），这工作得很好，但现在有几个“死”目录。 据我了解，HG 工作于文件而不是跟踪目录，所以我认为这就是正在发生的事情。
示例
原始目录结构：
/
- includes/
- bootstrap
- index
进行更改并推送（这是我在工作副本中看到的内容，它是最新的）：
/
- project/
- includes/
- bootstrap
- index
结果（这是我在远程存储库上看到的内容）：
/
- includes/
- project/
- includes/
- bootstrap
- index
这仅发生在远程存储库上，我的工作副本很好。 我已经看到 push 命令可以在远程存储库上运行远程命令 - 是否可以使用它来运行某些东西？
In my Hg repository (at Google Code) I have a directory that shouldn't be there. Basically I decided to move everything into a sub-directory (I had everything in the root), which worked fine but now there's a couple of 'dead' directories. Its my understanding HG works off files rather than tracking directories so I think that's what's happening.
Example
Original directory structure:
/
- includes/
- bootstrap
- index
Change made and pushed (this is what I see in my working copy, which is up-to-date):
/
- project/
- includes/
- bootstrap
- index
Result (this is what I see on the remote repo):
/
- includes/
- project/
- includes/
- bootstrap
- index
This only occurs on the remote repository, my working copy is fine. I've seen that the push command can run remote commands at the remote repo - could there be something I could run using that?
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（2）
我假设您正在谈论此cset（因为存储库是公开的，提供链接会更容易:)）
似乎只有谷歌代码显示是错误的，而不是你的存储库：查看此错误报告。
检查远程存储库中是否存在差异的一种简单方法是将存储库克隆到新目录中：新克隆包含完全相同相同的数据，并且其行为完全与main repo（这是一个 DCVS，因此每个节点都可以充当服务器或客户端，没有区别）。 因此，如果新的克隆是干净的并且按预期工作，您的远程存储库将处于完全相同的状态。
I assume you are talking of this cset (since the repo is public, providing the link would have been easier :) )
It seems that only the google code display is wrong, not your repository: see this bug report.
An easy way to check for discrepancies in your remote repo is simply to clone your repo in a new directory: a new clone contains exactly the same data, and behaves exactly as the main repo (this is a DCVS, so each node can act as a server or as a client, without distinction). For this reason, if a new clone is clean and working as expected, your remote repo is in the very same state.
谷歌自定义后端可能有问题吗？
Google 不为 code.google.com 使用标准的 Mercurial 后端。 我想知道该目录是否确实存在于 BigTable 后端中。 然而，在克隆时，mercurial 没有跟踪目录的概念，因此无法使用 Mercurial 协议传输信息。
如果您确实希望该目录消失，您可以执行存储库擦除，然后重新推送本地副本。
Could Google custom backend be at fault here?
Google doesn't use a standard mercurial back end for code.google.com. I wonder if the directory is acutally there in the BigTable backend. However when cloning, mercurial has no concept of tracked directories so that information can't be transmitted using mercurial protocol.
If you really want the directory gone, you could do a repo wipe then re-push your local copy.