Hacker News Reader

Comment by therein | Hacker News Reader

Would be nice if librewolf and WaterFox changed their flow to make it clear that Mozilla's Sync Server is being used, and make it opt in with something like:

To sync please select one of the options below:

[x] Custom Sync Server: ___________

[ ] Mozilla Sync

View on HN

LibreWolf is opt in and if you click the checkbox named "Firefox Sync" thinking you're getting some kind of sync that isn't Firefox Sync I don't think there's any UI change that can help you.

Your parent comment's idea is quite sensible, no need to be contemptuous like this. Firefox branding is a mess, being both a browser and a set of services. It's reasonable to mix things up or to be confused.

Sync is nothing but an awfully overengineered blob storage (I have implemented an alternative Accounts+Sync server when 1.5 came out, https://gitlab.com/drdaeman/firesync/ - long abandoned and the code is awful proof-of-concept inly, but the point here is to prove that I have a slightly educated opinion on the matter).

IMHO, It would be best (for technical users) if Firefox forks would rather replace Sync with something more standard like S3 or WebDAV. If my memory serves me right, there’s absolutely nothing in Sync that those protocols cannot do equally well (and even better because they’re actual mostly-interoperable standards).

Heck, it would be nice if Mozilla would do the same thing, but I’m afraid NIH is way too strong with them. I’m very happy if that gives us Rust, but Sync is simply garbage that shouldn’t have been ever made.

Is there a specification somewhere for the sync server that you followed, or did you figure out the endpoints/logic by looking at mozilla's sync server?

I was thinking about taking a crack at this as well, but I can't find a spec, and I want to leave looking at the python/rust sync server source code as a last resort.

The API docs are linked here: https://github.com/mozilla-services/syncstorage-rs?tab=readm...

It is documented here: https://mozilla-services.readthedocs.io/en/latest/storage/ap...

I'm not sure about the current status of this document. Years ago there were subtle nuances that I gad to reverse engineer looking at the source code and Firefox's actual behavior.

With only object storage, handling sync conflicts [0] becomes more complicated. I can understand why they don’t do that.

[0] https://mozilla-services.readthedocs.io/en/latest/storage/ap...

Yet stuff like CalDAV which is used by millions transparently handles syncing on top of WebDAV (arguably not just dumb storage , but not much better than it). Not only that, but Firefox Sync _itself_ used to be WebDAV.

For the record, for a lot of time Firefox Sync used to be exactly like that.

https://blog.mozilla.org/services/2011/03/22/get-syncing-wit...

Not only one of the first questions would be whether you want to use a Custom Sync Server, but this "Custom Sync Server" could be something as trivial as WebDAV server (if you have any kind of public-facing HTTP server, it almost certainly already supports this, or does so with minimal extra dependencies) . Then it would also ask you for the E2EE key, separately from the login key.

Reply on news.ycombinator.com