Welcome to logseq-sync Discussions!

[bcspragu]

Hello!

Figured I'd open up Discussions for things that aren't quite issues, feel free to put whatever here.

[Azarielh]

Hi !
Just find out this project whille looking for ways to sync my logseq work and since I've got a nas, I'm quiet interested about this project.
Doesn't see any activity since february. Is it still a thing?

[bcspragu]

Hello!

It's still a thing in the sense that I'm willing to work on it, but it seems we're at an impasse with the Logseq folks, who (perhaps understandably) don't seem to want to add support for non-official sync services.

[thouseef]

Do we then consider a fork of logseq if there is enough interest from the community?

The logseq folks are leaving a bad taste in our mouths with their treatment of your work. Being open source, we expect some level of open mindedness about engaging with the community. Feels like they feel threatened about this - thus ignoring anything related to the topic.

[megawron]

I think we should fork it. At some point if we are lucky the Logseq team may merge with our fork. But for now we waited already so long for any response

[bcspragu]

I agree with the sentiment, but I think forking Logseq would be tough, because:

1. Written in Clojure, which comparatively few people can (or are willing to) develop in
2. Large codebase and community behind the original (self-hosted sync is a fairly small hill to die on)
3. The codebase is in the middle of a ginormous persistence migration/refactoring (see the feat/db branch)

Of course, if folks are capable and willing to get a successful fork running, I'm happy to contribute the relevant changes to that fork (to add the settings, update rsapi, etc).

If I personally were going to do something, I'd indulge in a performance-focused ground-up non-Electron (i.e. closer to native) rewrite. I'd strip all the features I don't use, which would really just leave basic markdown/page/block/properties support, syncing, the query API, and a few others. I'd write it in Rust (in b4 RIIR), and use something like cozo for the persistence/Datalog layer.

But that's a pipedream for me, I don't really have the time (or even necessarily the desire) to do that.

A quick edit, since I just noticed this:

The logseq folks are leaving a bad taste in our mouths with their treatment of your work. Being open source, we expect some level of open mindedness about engaging with the community. Feels like they feel threatened about this - thus ignoring anything related to the topic.

I do disagree with this. Being open source, they are welcome to do whatever they want with their codebase and have no obligation to engage with me or anyone. I don't feel personally slighted or anything like that. Of course, I'd prefer if they just said "hey yeah we'll never support this don't bother" rather than radio silence, but I think both are perfectly acceptable responses. Like we're discussing, if the community doesn't like that, they have options (mostly just forking)