!NasysSDfxKxZBzJJoE:matrix.org

#matrix-spec

455 Members
Discussion of specific Matrix Spec Change (MSC) proposals and surrounding ideas | https://spec.matrix.org/unstable/proposals/ | SCT Board: https://github.com/orgs/matrix-org/projects/31 | Old design drafts: https://drive.google.com/drive/folders/0B4wHq8qP86r2ck15MHEwMmlNVUk 170 Servers

Load older messages


SenderMessageTime
3 Oct 2022
@deepbluev7:neko.devNicoA delete endpoint is useful, but how is it more flexible, if it doesn't even fullfill the original requirements?12:01:42
@kim:sosnowkadub.deHarHarLinks

i can either

  1. call patch with a diff
  2. file gets cloned
  3. maybe old file gets deleted implicitly
  4. return new mxc that i can edit in my event

or

  1. upload new/patched file completely
  2. return new mxc that i can edit in my event
  3. delete old file explicitly
12:03:36
@deepbluev7:neko.devNicoIf you upload the new file completely, you don't need a delete endpoint and the original question was explicitly about not uploading a new file12:04:33
@kim:sosnowkadub.deHarHarLinksi think patching makes little sense with e2ee12:04:55
@ze_test:matrix.orgBroccolo Romanesco joined the room.17:33:59
@mscbot:matrix.orgmscbot[MSC3902] "MSC3902: Faster remote room joins over federation (overview)" has been created: https://github.com/matrix-org/matrix-spec-proposals/pull/390218:31:53
4 Oct 2022
@michaelw:element.ioMichael Weimann changed their display name from Michael Weimann (back on 4/10) to Michael Weimann.06:44:37
@andrewm:element.ioAndrew Morgan (anoa)

Half-Shot: wondering if you can think of any use cases for including remote users in an AS' user namespace? re: https://github.com/matrix-org/synapse/pull/14000#issuecomment-1266552433

I was thinking remote user presence, but then that would likely be covered by listening to the presence updates that a local user receives

08:20:13
@Half-Shot:half-shot.ukHalf-ShotI can't really think of any. Local users lines up quite nicely with federation behaviour (i.e. in order to get events, you have to have your host in the room. In order to get events over the AS, you have to have a local user in the room). I can easily see people getting confused as to why they don't get a remote user's information down the AS api, and it turns out because they don't have any local users who share rooms with them etc.08:23:03
@Half-Shot:half-shot.ukHalf-ShotIt's ultimately more likely to be a footgun imo.08:23:14
@Half-Shot:half-shot.ukHalf-Shot The only nice thing I can see is you are being explicit about what users on the remote side you are interested in, but for me that's not a huge win. 08:24:11
@andrewm:element.ioAndrew Morgan (anoa)Yeah, especially as you'll still be receiving events from other remote users, which may be confusing.08:24:49
@andrewm:element.ioAndrew Morgan (anoa) cc tulir as well 08:25:48
@andrewm:element.ioAndrew Morgan (anoa)https://github.com/turt2live/matrix-monitor-bot/issues/45 argues for specifying remote users in the user namespace. But I believe having a regex for all local users would cover this use case..08:48:41
@Half-Shot:half-shot.ukHalf-Shot My feeling is that the AS API does need a filter API / config of some kind to filter out stuff you don't care about, which is logically seperate from namespaces imo. 09:39:54
@Half-Shot:half-shot.ukHalf-ShotWhich would help the use case of "I only care about these events, or these remote users".09:40:55
@Half-Shot:half-shot.ukHalf-ShotNamespaces are typically things you control / own in some fashion, and ASes can never control remote users obviously :)09:41:23
@Half-Shot:half-shot.ukHalf-Shot
In reply to @Half-Shot:half-shot.uk
My feeling is that the AS API does need a filter API / config of some kind to filter out stuff you don't care about, which is logically seperate from namespaces imo.
I should say: Largely the reason this doesn't exist is that it's far easier to simply /dev/null all the stuff you don't care about on the AS side. The only reason for it to exist is performance / privacy reasons.
09:42:58
@andrewm:element.ioAndrew Morgan (anoa)Sounds like we're in agreement about limiting what the namespace covers then - thanks!09:49:34
@andrewm:element.ioAndrew Morgan (anoa)(And yes, I came across your MSC https://github.com/matrix-org/matrix-spec-proposals/issues/3806 🙂)09:49:47
@Half-Shot:half-shot.ukHalf-ShotI was much younger and optimistic back then!09:55:19
@andrewm:element.ioAndrew Morgan (anoa)weren't we all 🧓09:55:47
@travis:t2l.ioTravisR
In reply to @andrewm:element.io

Half-Shot: wondering if you can think of any use cases for including remote users in an AS' user namespace? re: https://github.com/matrix-org/synapse/pull/14000#issuecomment-1266552433

I was thinking remote user presence, but then that would likely be covered by listening to the presence updates that a local user receives

There are T&S aspects which are useful, however these are also covered by an implicit .* rooms namespace
13:48:46
@condencionado:matrix.tosdr.orgcondencionadoVEVO changed their profile picture.17:08:52
@mscbot:matrix.orgmscbot[MSC3903] "MSC3903: X25519 Elliptic-curve Diffie-Hellman ephemeral for establishing secure channel between two Matrix clients" has been created: https://github.com/matrix-org/matrix-spec-proposals/pull/390317:33:29
5 Oct 2022
@naveen:teams.amrkernel.comNaveen Chandra joined the room.04:18:19
@bh7953:kit.eduFlorian Jacob joined the room.20:05:09
@florian:dsn.tm.kit.edu@florian:dsn.tm.kit.edu left the room.20:05:13
6 Oct 2022
@kerrya:element.ioKerry Archibald [out] changed their display name from Kerry Archibald to Kerry Archibald [out].11:31:12
@benp:element.iobenp (away) changed their display name from benp to benp (away).23:04:41

There are no newer messages yet.


Back to Room List