Lessons learned from building a sync-engine and reactivity system with SQLite
finkelstein.fr
finkelstein.fr
I work at Electric and started the PGlite and now Tanstack DB projects. The issues mentioned with PGlite are one of the major motivating factors behind Tanstack DB. We are taking those learnings and building, what we believe, is the missing client side datastore that is "sync native" and completely backend agnostic. Also being JS, rather than WASM, solves many of the slower than ideal query semantics, and has enabled us to build an incremental query engine for it.
It's also important to note that Electric doesn't require PGlite on the client, far from it - it's essentially a "protocol first" sync engine, you can use it to write into and maintain any client side store.
This solution by the OP, diffing based of modified data is ideal for a huge number of apps, and something that we intend to built into Tanstack DB so you can easily sync with no additional infrastructure.
SQLite (or PGlite) in the browser is awesome, and has the advantage over Tanstack DB at the moment of having persistence (it's on our roadmap), but they are also somewhat chunky downloads. For many local-first apps that's not a problem though.
It is opinionated and not for every use case, also very experimental, but you might find some of the ideas interesting.
Everything is a full new row because it is “a message” including read receipts. Some messages like read receipts just don’t render in the chat.
Edits can work the same way by rendering over a previous message, even though the local and remote DB have multiple rows for the original and edited messages.
I was very surprised (or perhaps disappointed is a better word) when I didn’t see Lamport, paxos or raft mentioned at all. At least crdts made an appearance, although almost in the post scriptum.
An interesting tool that matches the requirements mentioned in the article is Evolu[0]
It's a sync engine with e2e encryption based on SQLite.
The local-first landscape is quite wide now, and there is probably a solution ready for all kind of needs[1]
Building a sync engine can be a nice learning experience, but for production software it's better to pick something that has already faced and resolved all the weird edge cases you get when building a sync engine and persistent storage on the browser.
What does a document system need? A few simple indexes that track links and backlinks, mentions, and hashtags. You’re broadcasting change notifications anyway (for reactive DOM updates), so updating the indexes for those manually isn’t much extra work. You apply updates eagerly and then notify other users so they can apply the same updates, with some rules to guarantee that all users end up with the same state regardless of the order in which they receive the updates. But a relational database doesn’t help with any of this. Document systems tend to be versioned, so every user action turns into another entry in a transaction log. Even queries like “last Monday’s version of my document” don’t map naturally to SQL. You can query for transactions in a given time period, but unless you have snapshots, you’re still forced to re-apply transactions all the way from t=0 if you want to re-create the document state at a given date.
I don't see how this certainty follows from "granularity" (whatever that means in this context). I believe to have such a certainty one would need the synchronization to happen within a single transaction that spans both client and server databases.
Using the wa-sqlite library, and our own Mobx-based model layer similar to what Linear does.
I gave a short talk about it a few months ago https://www.youtube.com/watch?v=b0GKWzeyZko
Not so sure about this. These seem more like fundamentals than sliding scales.
> How many people will concurrently edit the same resources?
More than 1.
> How write-heavy is it?
Write-heavy enough that you'll encounter an unexpected write between two reads.
> Can you expect unreliable connections or offline usage?
Yes.
CAP with Electron and SQLite is no different from CAP with Tauri and MySQL.
Is it an issue with reactivity between browser tabs? Hence the use of the Broadcast Channel API?
The reactivity being implemented in svelte, what is different from any normal app? could have been react with mobx, zustand, recoil, etc but that's not too relevant. This is just how frontend frameworks are nowadays.
Especially since it is single-user (player? sic)?
Syncing between the backend server (remote state) and the local sqlite instance (local state) shouldn't really require much work? A simple websocket connection could do the trick to push from the remote server to the local device and then push the changes to sqlite again. But that's not too surprising to me.
I must not have understood something...
A lot of these tools (and what the author wrote) offer toolkits to do this well, not having to implement and track all these changes in state manually.
In this case the author is running SQLite in the browser, and that syncs to SQLite on the server.
ElectricSql, instantdb, rxdb, jazz.tools, are some of the things I’ve been looking into to make these tasks easier.
If someone shares their additional take on rewriting in rust (I say this with love as the type of person to do that) it doesn't add as much value because many of those experiences are shared regularly. If you share your unique tech stack, the people in the woodwork who might be struggling to do something along those lines but can't find the resources will benefit way more.
So you should!
Also can you elaborate on the CAN bus part of your smart home? Is your car wired into your home network?
It mirrors the moment I embraced tools like Obsidian; designing for real-world constraints; valuing simplicity, privacy, and functionality; especially in low‑connectivity environments.
PS: I was hooked on ...SQLite
Bravo!
Really people should just move to Gleam and Lustre or Elm, even if it means I'll have much fewer clients paying good money to untangle state issues...
> First Try: PGlite and Electric
Out of curiosity, apart from the lack of LISTEN support what more did you miss in SQLite itself and its ecosystem?