git clean removes content of symlinks #607
Comments
Please note that it is unreasonable to expect everybody to have
I hope you realize that a statement like this is prone to cause the opposite of what you want. After all, who likes to be underappreciated and pushed around? |
|
As far as I understand npm uses junctions (can be created by To reproduce the bug:
Now you will see that README.md is missing in target. |
|
EDIT: I messed this up. This comment and the next four are me accidentially leading this issue in the wrong direction.
Ummm... you have the folder target inside the repository, add the files inside it to the repository, and wonder that |
|
I don't think so, because executing |
How about |
|
Sure this will delete the target folder and the README.md, but this is not the point. My problem is that I have a folder "target" with files and obviously a junction "link" (a type of symlink) that links to this folder. Target and is files is wanted inside the repo and the junction "link" is not. Therefore, the the junction "link" is in the What I expected from |
Sorry. I messed this up by not looking up, what exactly Please accept my appology for wasting your time. I can confirm that |
|
No problem. I think with my addition it is now more precise. |
|
so basically you want |
|
Exactly. J Wyman [email protected] schrieb am Mi., 20. Jan. 2016 19:19:
|
I implemented similar logic for Libgit2 some time ago. I can assist here as well, though I doubt I'd have time to generate a full pull request on the topic. Win32's |
|
After such a long time, I have made up my mind. As pointed out in #437, junction points are just different enough from symlinks to not be treated as such. If you create a directory junction point, you really make the target directory accessible as a directory in the local directory. Note: this is not a symbolic link, but more like a "hard link" (if directories could be hard linked). In short: if you have a directory junction point and try to delete it recursively, I think that we do the exact right thing by removing the directory contents, too. @illuxio if you disagree, please try to convince me that this is an improper take. And be prepared to do the heavy-lifting of the change after that. |
|
@dscho git for windows should handle junctions the same way as other Windows programs do -- delete junction but not its content. I.e. when in Windows Explorer you delete directory that contains junction you delete just junction but not its content. The same way other Windows file managers behave. E.g. cmd.exe command |
|
@stasberkov I could be convinced by a patch implementing that logic ;-) |
|
Directory junctions are the most convenient softlinks for Windows NT directories — the symlink directory softlinks have lots of problems, like requiring special privileges, being interpreted at user level rather than by filesystem, which results in different processing depending on the client (e.g. when exposing the folder as a network share). So junctions are a sane choice. And we usually expect them to behave as softlinks in case of deletion. We experience undesired side-effects from the How about the implementation which isn't dependent on extra system calls (we use similar technique in ReSharper and Rider filesystem components to handle deletion consistently):
The second and third steps are already there, this just adds the first probe to kill the directory (or link source) directly. Haven't seen any noticeable side-effects from this so far. Is very easy to implement. Does not dig into filesystem implementation details. |
|
@hypersw I like your approach, it makes a lot of sense; however I have to say that there are already those that depend on the existing behavior and changing it could/would cause them heartburn and send them here complaining in the opposite direction. The best solution here would be for somebody who really wants and/or needs this functionality to offer up a solution via pull-requests which implements the solution; and places it behind a configuration flag so that users who depend on the existing functionality aren't broken by the improved functionality. Why do I suggest "somebody else" do the work? Personally, I'm already 110% scheduled and @dscho is likely in a similar situation. If those who want/need this change are happy waiting until such a time that somebody has available time to implement it, then that's fine - however if they're not happy to wait, then they need to act. Open-source is awesome like that. 😄 |
|
@hypersw Is the idea easy to code and test within the GfW codebase? A proof of concept (PoC) pull request (PR) may help things along. Does this make all the mechanisms (junctions, reparse points, etc) the same as the upstream Git's view of symlinks? And if not, how should it be documented so that it is really easy to understand (especially as 95% of web explanations would be the upstream viewpoint, so Windows users may never see the documentation...). |
|
Running into the same problem in the context of building a mulit-package repro with junctions set up by the NPM package manager. The behavior of I could find the time to try to tackle this, but I will probably need some help navigating this repro's conventions. @whoisj I have a hard time imagining how anyone might consider this behavior a feature. Can you cite an example? |
|
From: "Burt Harris" ***@***.***>
Sent: Sunday, December 24, 2017 12:44 AM
In tests (*** t1013-read-tree-submodule.sh ***) I'm getting messages like
`
not ok 4 - git read-tree -u -m --recurse-submodules: added submodule
removes an untracked ignored file # TODO known breakage`.
Does the TODO comment indicate these are known issues I can ignore in
testing my PR?
Pretty confident that this is the case, i.e. a known issue. Haven't dug to
find the particular test.
|
…nks] There was already code in the readlink function of mingw.c that correctly handled mount points (AKA 'junctions') so it was rather straight-forward to replicate that code into the two other locations where links are handled. I tested before and after these changes and git clean now handles junctions like other soft links.
…nks] There was already code in the readlink function of mingw.c that correctly handled mount points (AKA 'junctions') so it was rather straight-forward to replicate that code into the two other locations where links are handled. I tested before and after these changes and git clean now handles junctions like other soft links.
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
It seems to be not exactly rare on Windows to install NTFS junction points (the equivalent of "bind mounts" on Linux/Unix) in worktrees, e.g. to map some development tools into a subdirectory. In such a scenario, it is pretty horrible if `git clean -dfx` traverses into the mapped directory and starts to "clean up". Let's just not do that. Let's make sure before we traverse into a directory that it is not a mount point (or junction). This addresses #607 Signed-off-by: Johannes Schindelin <[email protected]>
Hallo,
I'm using git version 2.7.0. I discovered recently that a
git clean -xdfdeletes the content of my symlinks that are in my.gitignorefile instead of removing the symlinks as in unix.The symlinks was created by the command
npm link.I hop you will fix this soon.
The text was updated successfully, but these errors were encountered: