4 Dec 2024 |
olivia | I think what I'll probably go for do "bunch of individual state requests" for now but try to leave space for adding filtered sync or SSS alternatives later | 07:23:24 |
olivia | the issues with a non-sync implementation that I've found so far don't affect migrating my own user, so I can get that out of the way and then come back to it | 07:24:02 |
Emma [it/its] | Tbh my main issue with lack of filtered sync is payload size | 07:30:31 |
Emma [it/its] | But it's something i account for sometimes in MatrixUtils (aka running multiple sync loops for synapse, one for conduit, just based on whether filtering works) | 07:31:23 |
olivia | In reply to@emma:rory.gay Tbh my main issue with lack of filtered sync is payload size hmm, just on bandwidth use or you think the payload size is responsible for significant latency on it's own? | 07:36:08 |
Emma [it/its] | For me its mainly bandwidth | 07:42:27 |
Emma [it/its] | Yknow, 12mbps connection and all | 07:42:43 |
Emma [it/its] | My synapse init syncs kn seconds, but thats all void when it takes me a minute to download | 07:43:46 |
olivia | In reply to@emma:rory.gay Yknow, 12mbps connection and all ah yeah, that makes sense | 07:46:21 |
olivia | In reply to@emma:rory.gay My synapse init syncs kn seconds, but thats all void when it takes me a minute to download initial sync with my account on computer.surgery grapevine takes like multiple minutes. I haven't actually checked how much of this is transfer time, but I think not most of it | 07:47:16 |
olivia | I suspect this is because the server-side sync code is fucked lol | 07:47:23 |
Emma [it/its] | Thats odd | 07:47:29 |
olivia | it's good to know that synapse initial sync is much faster, so it's not fundamental | 07:47:55 |
Emma [it/its] | okay i was trying to time and it staking a while | 08:01:52 |
Emma [it/its] | In reply to @olivia:envs.net it's good to know that synapse initial sync is much faster, so it's not fundamental apparently i was wrong | 08:02:34 |
Emma [it/its] | so far 1m30 without response | 08:02:56 |
Emma [it/its] | % Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 160 100 160 0 0 1 0 0:02:40 0:02:00 0:00:40 45
curl 'https://matrix.rory.gay/_matrix/client/v3/sync?timeout=0' -H -H -H - 0.05s user 0.09s system 0% cpu 2:00.54 total
| 08:04:05 |
Emma [it/its] | oddly small response | 08:04:43 |
Emma [it/its] | % Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 48.3M 0 48.3M 0 0 2920k 0 --:--:-- 0:00:16 --:--:-- 8034k
curl 'https://matrix.rory.gay/_matrix/client/v3/sync?timeout=0' -H -H -H - 0.29s user 0.31s system 3% cpu 16.975 total
there it is | 08:04:55 |
kay0z | Hi guys,
has anyone ever thought about how to bridge matrix to matrix? It sounds like a dog bite its own tail. :-D But seems to be necessary for my use case, at least out of pure interest | 14:22:49 |
tulir | there are a bunch of people who want that so it's not a new idea, but none of the people who want that have the time and skill to build such a bridge | 14:24:54 |
ninchuka | I've heard/seen people talk about using one they made but never saw a repo for it sadly | 14:26:04 |
kay0z | Im one of those people :-) | 14:31:26 |
kay0z | patience is a virtue :-) | 14:31:49 |
| Freezlex changed their profile picture. | 14:37:42 |
David Blackman | I came here from the synapse developer docs, I believe that this room is matrix dev in general, not synapse dev specifically, so forgive me if this question is in the wrong place.
I'm trying to run complement tests against my synapse fork in github actions. When I do so, a bunch of tests fail remotely that don't fail locally using the complement.sh script in the synapse repo. Most of the tests fail on Bad JSON value: float . For anyone who has worked with complement tests before - is there any obvious reason why this might be happening?
| 15:29:31 |
richvdh | there's a #complement:matrix.org room where you might get some help | 15:31:13 |
| ☂ ニクス Natsu ☂ (loli fan, the geometry dash player, English cuz yes~ ང་ཧོ་ཤི་ནོ་ལ་ཧ་ཅང་དགའ། ཨོ་ངའི་དཀོན་མཆོག མོ་ནི་ཡིག་ཚགས་སྔོན་པོ་དེ་ཧ་ཅང་མཛེས་མོ་འདུག དགའ་སྤྲོ་ཆེན་པོ་འདུག ཨོམ་ང་ཧོ་ཤི་ནོ་ལ་ཧ་ཅང་དགའ་པོ་ཡོད།) changed their display name from ☂ ニクス Natsu ☂ to ☂ ニクス Natsu ☂ (loli fan, the geometry dash player, English cuz yes~ ང་ཧོ་ཤི་ནོ་ལ་ཧ་ཅང་དགའ། ཨོ་ངའི་དཀོན་མཆོག མོ་ནི་ཡིག་ཚགས་སྔོན་པོ་དེ་ཧ་ཅང་མཛེས་མོ་འདུག དགའ་སྤྲོ་ཆེན་པོ་འདུག ཨོམ་ང་ཧོ་ཤི་ནོ་ལ་ཧ་ཅང་དགའ་པོ་ཡོད།). | 15:31:30 |
David Blackman | rad thanks, I couldn't find it earlier | 15:31:08 |
| Administrator banned ☂ ニクス Natsu ☂ (loli fan, the geometry dash player, English cuz yes~ ང་ཧོ་ཤི་ནོ་ལ་ཧ་ཅང་དགའ། ཨོ་ངའི་དཀོན་མཆོག མོ་ནི་ཡིག་ཚགས་སྔོན་པོ་དེ་ཧ་ཅང་མཛེས་མོ་འདུག དགའ་སྤྲོ་ཆེན་པོ་འདུག ཨོམ་ང་ཧོ་ཤི་ནོ་ལ་ཧ་ཅང་དགའ་པོ་ཡོད།) (expressing support for CSEA). | 15:41:48 |