Hacker News Reader

Comment by rw_grim | Hacker News Reader

We're finally gearing up to have an experimental release of Pidgin 3.0 by the end of the year, but the goal right now only include the IRC protocol. But everything has been updated to support all of the newer chat features so support for other protocols should come quick.

View on HN

(warning, heavily opinionated post follows)

I know it's asking for a lot, but it would be really cool if Pidgin would have 1st-class out-of-the-box support for Matrix.

I don't want to get into discussions if it's better than Jabber, because I don't really think it is, but since the momentum is on Matrix rather than XMPP, then I'd say that Pidgin could use the fact that currently Matrix lacks a proper client. By "proper client" I mean something that is feature-complete by standards of year 2000 (actually good software, like Pidgin), not 2020 (which features broken, half-ass web prototypes that people call software).

It would probably help with fighting the parasites like Discord, which is way too popular than it should be.

someone could certainly pick up https://github.com/matrix-org/purple-matrix and finish it.

The current state of purple-matrix for use in Pidgin leaves a lot to be desired. For example, it's quite slow to connect and missing a decent amount of features which aren't just nice to have. OTOH, the format of chats is a bit more streamlined and clearer to read.

Here's hoping the next version of Pidgin implements something that resolves the slow connection so I can begin using Pidgin as my preferred Matrix client over Fractal or the like.

That plugin was abandoned and we (the pidgin team) are unlikely to pick it up as we're planning our own implementation.

We have been planning a new from scratch version that'll be in tree, but with the retirement of libolm which is for good reasons, it means we're going to have to write our own OLM implementation at some point as well.

https://matrix.org/blog/2024/08/libolm-deprecation/

TL;DR it's on the list, going to be a bit before we get to it.

hm, you really shouldn’t have to write your own olm implementation(!)

either you could swap primitives in libolm (eg fork libolm and merge https://gitlab.matrix.org/matrix-org/olm/-/merge_requests/24) or use vodozemac via wrappers.

libolm was not using battle tested crypto. That's one of their main reasons for abandoning it. Our plan is to use gcrypto for it which is battle tested.

As far as vodozemac goes, we're not pulling rust into our build system.

Reply on news.ycombinator.com