Skip to content

Native iOS port / GUI #102

Closed
Closed

Description

@sloev

Developing an ios client for syncthing.
right now GO doesnt easily compile using xcode but is compiles manually on-device (requires jailbreak)

What will it take to make an ios client?
should it embed the GO syncthing or be a complete rewrite specific for obj-c?

lets discuss!


This issue is closed for discussion. Vote or add bounties etc here: https://www.bountysource.com/issues/7699463-native-ios-port-gui

// @calmh 2016-08-18

Activity

  1. calmh commented on Mar 24, 2014

    @calmh
    Member

    That's an awesome initiative!

    I don't think it's possible to get an app approved that contains foreign binaries (such as a standalone syncthing written in Go). Go currently doesn't support being linked as a library to other C (or Objective-C) code, due to the runtime. So as it is, I think the only way forward that is guaranteed to succeed on that front is a pure Objective-C implementation.

  2. sloev commented on Mar 24, 2014

    @sloev
    Author

    it would be AWESOME(!!! repeat forever) if you guys included an example of implementing the protocol in another language.

    aside yes i guess implementing the protocol in pure obj-c is the way however it seems i have nowhere to start (since lack of third-language implementation).

    i am willing to make a big effort on the ios side but the objective would be to keep inside the frames of non-jailbreaked phones and free of course :-)

    again "third party" implementation of protocol/functionality is highly needed

  3. calmh commented on Mar 24, 2014

    @calmh
    Member

    Someone has to be the first third party, and it obviously can't be me since I'm the first party. :)

    I would start with reading https://github.com/calmh/syncthing/blob/master/protocol/PROTOCOL.md and filing bugs on any ambiguities or other strange stuff.

  4. sloev commented on Mar 24, 2014

    @sloev
    Author

    is there similar docs (have searched but not reaaaaly found any) for the network/transport layers.
    i mean documentation on the establishment of lan/internet wide connections.

  5. filoozom commented on Mar 24, 2014

    @filoozom
    Contributor
  6. sloev commented on Mar 24, 2014

    @sloev
    Author

    yes thanx.
    i know its not good githubing to mention issues in issues in .... but that site is not listed from the wiki. :-)

  7. calmh commented on Mar 24, 2014

    @calmh
    Member

    Yeah... Beware that while I always intended the block exchange protocol to be a stable standard, I haven't put nearly as much thought into the discovery protocol. That was always more of a half baked lets-do-something-and-see-what-happens thing that might lead into a DHT or mDNS or something in the long run.

  8. calmh commented on Mar 24, 2014

    @calmh
    Member

    Also a note on the BEP protocol is that the specs kinda vaguely say something about it typically being used with TLS over TCP. I'm considering mandating that or interoperability will be tricky in the future. From the beginning I was considering running it over UDP etc but I'm not sure there is a legitimate use case for anything not TLS/TCP.

  9. sloev commented on Mar 24, 2014

    @sloev
    Author

    yeah i guess an iphone app with "just" the ability to show an index of files using a predefined ip is a starter.
    however if the discovery is not soon ot be set in stone then i guess it really disturbes development of third party stuff :-)

  10. calmh commented on Mar 24, 2014

    @calmh
    Member

    Agreed. Lets have a discussion about that when the first third party project is at the point where it's relevant. :)

  11. calmh commented on Mar 24, 2014

    @calmh
    Member

    About the discovery specifically, I really feel that is an area where a single app like syncthing should not have to make up it's own thing. There should be a standard ready to use. I think that standard is mDNS w. DNS-SD and that is where I'd like to go. I just haven't had the time / energy / prioritization to go there when a custom thing was easier to get off the gorund.

  12. sloev commented on Mar 24, 2014

    @sloev
    Author

    mDNS / dns-sd is sweet and extremely transparent in that it delivers knowledge on its services and optionally hints on how to use them.

    but it only works locally. you still need a standards solution globally.

  13. sloev commented on Mar 24, 2014

    @sloev
    Author

    by reading:
    https://github.com/calmh/syncthing/wiki/Case-Study
    i got a better understanding of the global lookup in this case using your announce server. is the announce server sourcecode available?

  14. 60 remaining items

  15. calmh commented on Dec 26, 2021

    @calmh
    Member

    I don't think any of the usual suspects are in any way interested in doing this, and be assured that doing it properly is a major piece of work. So I don't think there's anything to revisit, really.

  16. zjDrummond commented on Dec 27, 2021

    @zjDrummond
  17. calmh commented on Dec 27, 2021

    @calmh
    Member

    If it didn't fit our goals the issue would be rejected. Still needs someone willing and able to do it.

  18. DivineDominion commented on Sep 13, 2022

    @DivineDominion

    Yes, I've done this - Syncthing can run just fine on iOS when compiled to a C library and started by a thin wrapper application that provides a web view for the GUI. But it's not a very useful thing.

    @calmh Could you document how to compile into a C library? I'm a capable enough Swift developer to add the module map myself and export this as a proper Swift package, but I'm no good at understanding the build system of Go :)

  19. calmh commented on Sep 13, 2022

    @calmh
    Member

    It's been many years, I don't remember. I'm sure the info is out there, though.

  20. fuzihaofzh commented on Sep 28, 2022

    @fuzihaofzh

    It's 2022 now. Seems it still not not come out. I just wonder resilio also has a ios version. Is there any particular reason that syncthing still not support ios?

  21. bezmi commented on Sep 28, 2022

    @bezmi

    If it didn't fit our goals the issue would be rejected. Still needs someone willing and able to do it.

    I think it's pretty clear why. People are expecting an iOS port to materialise out of thin air. It's an immense amount of work and no one is willing to volunteer the time needed to do it.

  22. OliverCole commented on Sep 28, 2022

    @OliverCole

    The $515 bounty obviously isn't sufficient. Perhaps the issue needs to be broken down into smaller pieces people might be more able to help fund in the near term?

  23. pySec710 commented on Oct 1, 2022

    @pySec710

    What about creating a mobile web client and then making it an iOS app by embedding it in a frame? It wouldn’t interact with your file system obviously so no syncing but it would allow you to view the file system, down files, and maybe even add a built in file viewer like google drive. That way you get some interaction with syncthing and could be a good start to a full iOS client. It would also give users the option of a google drive style way to access their files without needing a client app installed.

  24. ivankovnatsky commented on Oct 3, 2022

    @ivankovnatsky

    I think this is what: https://github.com/MobiusSync/MobiusSync closely to what you suggest does, I guess.

  25. locked and limited conversation to collaborators on Oct 3, 2022
  26. imsodin commented on Oct 3, 2022

    @imsodin
    Member

    I locked the issue as the intent here is clear. More comments that it's needed or how it might be done don't help. Someone "just" needs to do work. For any discussions or even better, if you'd like to work on it, reach out on https://forum.syncthing.net.

  27. added theissue type on May 30, 2025
  28. calmh commented on May 30, 2025

    @calmh
    Member

    This now exists in the form of Synctrain 🎉

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    not-our-bugUsed only when we want to keep an issue open for visibility, but the real problem is somewhere else

    Type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions