24 May 2022 |
richvdh | In reply to @clokep:matrix.org (I really need a mapping between GitHub and MXIDs. âšī¸ ) @deepbluev7 on github is @deepbluev7:neko.dev , aka Nico, the maintainer of Nheko. He's not yet joined the new #synapse-dev. | 17:04:31 |
clokep | Thank you! | 17:04:45 |
clokep | I was pretty sure but didn't want to bing random people. | 17:04:57 |
Olivier 'reivilibre' | At some point, the cache eviction metrics started storing the reason for eviction. Is that in our recommended Grafana dashboard and if it is, maybe that would give us a clue? | 17:06:06 |
Olivier 'reivilibre' | seems like it is | 17:06:34 |
Olivier 'reivilibre' | (https://github.com/matrix-org/synapse/commit/a10988983a1cd145fc5ae57c9a00ea95fbaece61) | 17:06:43 |
clokep | See https://github.com/matrix-org/synapse/pull/12859 | 17:07:24 |
richvdh | In reply to @clokep:matrix.org I think it is waiting for the author to pick back up, but I can do the work if needed. another question on this: if you have work that depends on #11804, does that mean you've tried out #11804? One of my concerns is that maybe nobody other than nico has tried it, so if you have, that would be good information | 17:08:07 |
Github | https://github.com/matrix-org/synapse/pull/11804 : Implementation for MSC3664: Pushrules for relations | 17:08:09 |
clokep | In reply to @richvdh:sw1v.org another question on this: if you have work that depends on #11804, does that mean you've tried out #11804? One of my concerns is that maybe nobody other than nico has tried it, so if you have, that would be good information I have not tried it. I looked at the implementation a long time ago and it seemed reasonable. | 17:08:38 |
Github | https://github.com/matrix-org/synapse/pull/11804 : Implementation for MSC3664: Pushrules for relations | 17:08:38 |
richvdh | the trouble with push rules is that the server-side impl always seems reasonable | 17:09:08 |
tulir | In reply to @reivilibre.element:librepush.net At some point, the cache eviction metrics started storing the reason for eviction. Is that in our recommended Grafana dashboard and if it is, maybe that would give us a clue? it says (size) | 17:09:08 |
richvdh | it's when it starts integrating with clients that the wheels fall off | 17:09:18 |
clokep | Hm, alright. âšī¸ | 17:12:17 |
clokep | I will give it a try then...and poke Nico about whether they plan to update that or not. | 17:12:44 |
clokep | I imagine at least fixing the conflicts is easy enough. | 17:12:51 |
Olivier 'reivilibre' | In reply to @tulir:maunium.net it says (size) hmm, I don't have any futher ideas at the moment though :| | 17:15:03 |
tulir | hmm, maybe there's something wrong with the environment variable, seems fine now that I set it via config | 17:21:24 |
| madlittlemods (Eric Eastwood) joined the room. | 18:06:34 |
Nick Mills-Barrett | Hey all is anyone aware of anything preventing enabling the account/whoami endpoint for workers? I noticed it's not included in the generic worker servelets currently | 18:26:29 |
Nick Mills-Barrett | looking through the codepath for the API, it seems like it should be fine (basically just calls out to get auth from req) | 18:26:55 |
Andrew Morgan (anoa) | Nick Mills-Barrett: I agree that it's likely to be fine to run on workers. Would you be willing to PR that documentation change? đ | 18:28:52 |
Nick Mills-Barrett | In reply to @andrewm:element.io Nick Mills-Barrett: I agree that it's likely to be fine to run on workers. Would you be willing to PR that documentation change? đ Yep absolutely, will get a PR for the enable & doc tomorrow AM | 18:29:33 |
| Brad Murray joined the room. | 18:29:36 |
Andrew Morgan (anoa) | Cheers! | 18:29:39 |
clokep | We usually update the mapping in sytest (and complement) with that change too so that it gets tested. đ | 18:35:22 |
Brad Murray | In reply to @richvdh:sw1v.org another question on this: if you have work that depends on #11804, does that mean you've tried out #11804? One of my concerns is that maybe nobody other than nico has tried it, so if you have, that would be good information fwiw we're running the original code in production for a couple months with some good success, but that's not the same implementation as the one that got MSC'd and PR'd to synapse | 18:48:05 |
Github | https://github.com/matrix-org/synapse/pull/11804 : Implementation for MSC3664: Pushrules for relations | 18:48:05 |
Brad Murray | So I can speak to the general approach working for our narrow use case (selectively notifying on reactions depending on the author of the original message) but not the code in the PR being bug-free | 18:49:23 |