Feature/sync unix symlinks as links

[taminob]

Summary

This PR allows the synchronization of Unix symlinks on Unix-like systems. The symlinks will be synchronized as links. I try to provide an MVP implementation which, in the future, can be further extended to improve usability and add more features.

TODO

Features:

• [x] Allow upload of symlinks via single-file upload (PUT request)
• [x] Allow upload of symlinks via bulk upload (POST request)
• [x] Allow download of symlinks
• [x] Fix server querying which assumes that everything, that is not a directory, is a file
• [x] GUI option to enable/disable symlink synchronization
  • [ ] Hide option on Windows (if not compatible, see next point)
• [ ] Check Windows compatibility (current implementation might work for Windows shortcut files (except probably setModTime and readlink) since Qt handles shortcut files (.lnk) as Windows symlinks)

Known bugs:

• [ ] When enabling symlink synchronization with already existing symlinks, the info message "Files from the ignore list as well as symlinks are not synced." does not disappear until restart
• [x] Conflict with symlink will show size of 0 for broken symlinks or the size of the symlink target for valid symlinks
• [x] When downloading a regular file which overwrites a broken symlink, only the write permission for owner is set and the file is not readable
• [x] Conflict between two (broken?) symlinks different timestamps does result in message "Files from ignore list as well as symbolic links are not synced." for conflict file (although symlink synchronization is enabled)
• [ ] No conflict/synchronization for two (broken?) symlinks with same length and modification time (does also happen for regular files, probably shouldn't be addressed in this PR)

Tasks:

• [ ] Unit testing
• [ ] Regenerate translation files
• [ ] Rewrite git history

Related

Resolves #250 (and probably also #5509).

This change requires https://github.com/nextcloud/server/pull/41321 for the nextcloud server.

Disclaimer

This PR is still in a WIP state, do not test this on a production Nextcloud server. Also, everything in there is still subject to change and hacky solutions and code full with TODOs will appear in this PR.

[nextcloud-desktop-bot]

AppImage file: nextcloud-PR-6205-000987f9111afe85b85875b83d37ce6b99699017-x86_64.AppImage

To test this change/fix you can simply download above AppImage file and test it.

Please make sure to quit your existing Nextcloud app and backup your data.

[taminob]

@claucambra I planned to mainly ignore non-Unix systems for now and only ensure that it doesn't break the build and current behavior on those systems. I think that for now it's probably best to just not download symlinks from the server if on a Windows system. There have been already lengthy discussions around that in the issue and I think it would be best to handle it as a separate feature to slowly make progress regarding symlinks and then continue improving it step by step.

Could be that Windows will be relatively easy due to NTFS symlinks (I have no experience with them) and we only have to provide custom "symlink" and "readlink" implementations. Alternatively, Qt handles Windows "shortcuts" as the equivalent to symlinks as far as I can see in the documentation.

Also, I'll look into including an option to toggle this feature and (at least for an experimental phase) default it to off so that the current behavior will be kept as-is.

[AkechiShiro]

Any way we can help move this PR forward so it lands?

[taminob]

Any way we can help move this PR forward so it lands?

Since this PR depends on https://github.com/nextcloud/server/pull/41321 and the server maintainers are actively ignoring it, I don't think it makes sense to spend any time on this PR here. I can't think of anything other than upvoting the other PR and pinging the reviewers there.

[blob42]

Hi,

Is someone aware of a fork of nextcloud with this feature merged ? If someone would do it and setup some Patreon/donation I would gladly pay for this feature and I am sure many others will

[f1d094]

@taminob - It looks like the server-side portion was quietly approved/merged into 29.0.0 beta 2

[AkechiShiro]

Thanks for mentioning it @f1d094 I wasn't really sure given the silence from the maintainers side.

[taminob]

Hi @claucambra, on server side the proposal was rejected on the basis that the desktop app does not support an apps system to make the symlink support optional. However, I think that it should be possible to extend the desktop app such that the symlink support is in the code base, but can be toggled by the user similar to how it is done for the end-to-end encryption support (or also external storage). What do you think? Is this for you actually a major blocker or shouldn't it be possible to find a solution for that?

[taminob]

Hi @claucambra, on server side the proposal was rejected on the basis that the desktop app does not support an apps system to make the symlink support optional. However, I think that it should be possible to extend the desktop app such that the symlink support is in the code base, but can be toggled by the user similar to how it is done for the end-to-end encryption support (or also external storage). What do you think? Is this for you actually a major blocker or shouldn't it be possible to find a solution for that?

@claucambra ping

[f1d094]

@claucambra Claudio, after a few months away I received a few pings from other threads and it was brought to my attention that the core-holdup is here. I would like to share with you a cross-post I made for the server team:

First...I should point out that the lack of proper symlink synchronization is already causing a lot of issues and should be considered a bug, not a feature. As currently implemented Nextcloud actively breaks multi-client Linux/Mac Nextcloud setups. Secondly...I'd like to stress that if implemented correctly, it should be invisible to Windows users and transparently supported on Linux. Allow me to make the case:

1. On the Nextcloud server, there should exist a special filetype, that filetype is "POSIX-SymLink" or whatever.
2. The special filetype is ignored by all Windows clients. Full stop.
3. On Mac/Linux systems, it is sent client-side as a symlink. It doesn't matter where it points to, it could be a broken link, or it could point to /etc/shadow...that is not the concern of the client. A link is a link is a link. I can point my links to the moon and if there is no path there, what is the harm? I've just created a broken symlink. The client should never-ever resolve the link or care where it points. It is just a basic filetype.
4. Deleting the file on the nextcloud server, deletes the symlinks on clients that sync the symlink, as it would any other ordinary file.

Why is this at all complicated? Can you provide a use-case where the above would break any given install or cause comlications? Please explain.

We, as a team of Mac and Linux users, are incredibly frustrated with Nextcloud because of this very broken aspect of the way Nextcloud handles POSIX filesystems. If you claim to support Linux as a client, symlinks must be properly supported. They are a critcial and integral part of computing-on-Linux.

[gschenck]

Symlinks under Windows are no problem at all with NTFS as the standard file system.

[f1d094]

Symlinks under Windows are no problem at all with NTFS as the standard file system.

I have not used windows in a long time but it was my understanding that symlinks were not enabled by default. Nextcloud can neither anticipate nor require that users have them enabled in order for their client to function. Therefore, they should likely only be an option for Windows. Unless of course this has changed and they are enabled by default now on any version of the operating system currently supported.

[gschenck]

Symlinks under Windows are no problem at all with NTFS as the standard file system.

I have not used windows in a long time but it was my understanding that symlinks were not enabled by default.

https://learn.microsoft.com/en-us/previous-versions/windows/it-pro/windows-server-2008-R2-and-2008/cc753194(v=ws.10)

Symbolic links under windows have been available and used for a long time.

[claucambra]

Hi @claucambra, on server side the proposal was rejected on the basis that the desktop app does not support an apps system to make the symlink support optional. However, I think that it should be possible to extend the desktop app such that the symlink support is in the code base, but can be toggled by the user similar to how it is done for the end-to-end encryption support (or also external storage). What do you think? Is this for you actually a major blocker or shouldn't it be possible to find a solution for that?

@claucambra ping

Hi @taminob apologies for the delayed response.

We have discussed this in the desktop team and for us the key holdup in having this incorporated is the lack of automated testing for the changes introduced. If we have some auto tests we are happy to include these changes. From my understanding symlinks can just be synced as innocuous text files on Windows

[taminob]

Hi @taminob apologies for the delayed response.

We have discussed this in the desktop team and for us the key holdup in having this incorporated is the lack of automated testing for the changes introduced. If we have some auto tests we are happy to include these changes. From my understanding symlinks can just be synced as innocuous text files on Windows

Hi @claucambra, thanks for your response and for discussing this in the team. That is great news. :)

The current state of the PR is far from ready and I'd definitely add tests for the proposed changes before we merge it. At some point in development after I finished the core functionality, I just stopped finalizing the PR because I didn't feel like it would get merged at all anymore.

Unfortunately, I currently don't have any spare time to finalize this PR and do the necessary adaptions for the server. As I've written in https://github.com/nextcloud/desktop/issues/250#issuecomment-2400385724, I don't have too much time for the next ~10 months. Maybe I'll find some time in March to get started, but I'm not sure if I can wrap everything up then.

Once I have time (or anyone else wants to start earlier), the necessary steps will be:

• transform current server-core implementation to a server app
• write tests for this PR (after updating to latest master and resolving conflicts)
• test the changes of this PR on Windows (and also verify that it still works on Linux)