| 21 Jul 2021 |
 houston | I see | 05:38:39 |
| houston | if it did cache messages, then honouring redaction or not would make a difference | 05:39:17 |
| houston | thanks for all the info, I did post a bit of a wall of text there | 05:40:05 |
| houston | looks like alt/layer 4/etc also affect that scrolling thing, I guess that would be why it was different | 05:41:27 |
 kitsune | Even if it cached messages it would update the redacted pieces as soon as it receives redactions from the server. | 05:46:16 |
| kitsune | And yes, Alt is fine, too - it's just plain and Shift that go to the message entry rather than to the timeline (and the idea is to probably swap these because you normally don't need to page up/down the message entry box often. | 05:47:54 |
| houston | lame | 06:30:10 |
 Github [@kitsune:matrix.org] | [quotient-im/Quaternion] KitsuneRal opened issue #782: Last timeline position is not accurately saved [open] - https://github.com/quotient-im/Quaternion/issues/782 | 06:37:30 |
| kitsune | houston: you mean the redactions? Again, Quaternion may not even get the original message from the server. If you really want to sneak into pre-redacted messages, run your own homeserver. | 07:12:36 |
| houston | not that part, I thought you were saying that even if message caching was built into quaternion, it'll still delete any client-side messages that got a redaction flag | 07:47:04 |
| kitsune | That’s correct. Generally, it is discouraged in Matrix to preserve the pre-redacted text for anything but administrative purposes. | 08:20:44 |
|  teststststs joined the room. | 11:25:33 |
| 22 Jul 2021 |
|  Юля, якій сьогодні впадлу changed their display name from bodqhrohro to Юля, якій сьогодні впадлу. | 16:16:45 |
| Github [@kitsune:matrix.org] | [quotient-im/libQuotient] KitsuneRal opened issue #484: Initial support for Spaces [open] - https://github.com/quotient-im/libQuotient/issues/484 | 21:17:14 |
| Github [@kitsune:matrix.org] | [quotient-im/libQuotient] VarLad opened issue #485: Any planned Qt6 shift? [open] - https://github.com/quotient-im/libQuotient/issues/485 | 22:47:41 |
| Github [@kitsune:matrix.org] | [quotient-im/libQuotient] CarlSchwan closed issue #485: Any planned Qt6 shift? [closed] - https://github.com/quotient-im/libQuotient/issues/485 | 22:50:49 |
| 23 Jul 2021 |
| Github [@kitsune:matrix.org] | [quotient-im/libQuotient] KitsuneRal opened issue #486: Release 0.6.8 [open] - https://github.com/quotient-im/libQuotient/issues/486 | 06:41:25 |
| kitsune changed the room topic to "Quaternion 0.0.95 will require Qt 5.11 |
Screenshot wanted for upcoming 0.0.95 release | libQuotient 0.6.8 is coming |
CI binaries are at https://cloudsmith.io/~quotient/repos/quaternion |
Quaternion status: https://github.com/quotient-im/Quaternion/projects/2 |
lib status: https://github.com/quotient-im/libQuotient/projects/2" from "Quaternion 0.0.95 RC2! |
Screenshot wanted for upcoming 0.0.95 release |
CI binaries are at https://cloudsmith.io/~quotient/repos/quaternion |
Quaternion status: https://github.com/quotient-im/Quaternion/projects/2 |
lib status: https://github.com/quotient-im/libQuotient/projects/2". | 06:44:27 |
| 24 Jul 2021 |
| * kitsune just checked out the savings on using CBOR instead of JSON on a bigger room - 10% | 16:17:31 |
| kitsune | The amount of text is too large for CBOR to shine. | 16:19:05 |
 nephele | Just add more overhead to the events, then CBOR savings will increase ;) | 17:28:22 |
| kitsune | Overhead should be numeric though. If it's more text (more textual attribute values, e.g.), the percentage of savings will even decrease. | 19:55:12 |
|  diaspora joined the room. | 21:07:13 |
| 25 Jul 2021 |
|  tluvesuah left the room. | 01:15:13 |
| nephele | > Overhead should be numeric though. If it's more text (more textual attribute values, e.g.), the percentage of savings will even decrease.
Hmm, just base64 encode everything, then we can get 3/4 savings already | 11:04:02 |
| kitsune | What does CBOR have to do with Base64? | 18:02:09 |
| 26 Jul 2021 |
| nephele | you can un-base64 stuff because CBOR supports binary data | 08:02:53 |
| kitsune | Ah, that yes. But we don't have that much encoded for now - a lot of the payload is service information in Latin-1. That being said... | 09:34:44 |
| kitsune | ...messages in scripts that use a lot of non-Latin1 suffer even greater than base64-encoded payloads :( | 10:05:49 |
|  uhoreg changed their display name from uhoreg (away) to uhoreg. | 14:08:33 |
