!XxiikwJisobPZKqttb:matrix.org

#matrix-spec-process:matrix.org

44 Members
For meta-discussion about the MSC proposal process | https://matrix.org/blog/2018/06/20/towards-open-governance-for-matrix-org/ | https://github.com/matrix-org/matrix-doc/issues/131822 Servers

Load older messages


SenderMessageTime
6 May 2021
@travis:t2l.ioTravisRwithin reason, sure. Unstable prefixes don't go away after 2 months: some are forever part of the fabric of Matrix. It's up the individual implementations to decide if they care.21:21:05
@travis:t2l.ioTravisR For example, it's completely reasonable that Element gives up on org.matrix.msc1772.type (as was originally proposed, and still is). It's also completely reasonable if they decide that they'll check for org.matrix.msc1772.type and type. 21:22:55
@shadowjonathan:matrix.org[Jonathan]With “unstable prefixes” there I meant online endpoints, and emitting unstable identifiers, ofc past events are gonna keep those prefixes, I know that21:24:01
@shadowjonathan:matrix.org[Jonathan] * Offered endpoints, and emitting unstable identifiers, ofc past events are gonna keep those prefixes, I know that 21:24:10
@travis:t2l.ioTravisRin practice it's a very careful dance of how something becomes de-prefixed, usually involving multiple engineering teams. The server needs to support the stable endpoint first, then the clients need to check the server's support of that endpoint, then once sufficient adoption has happened the server can remove the unstable version and clients can stop trying the unstable version. The migration doesn't have to happen like this though, and whatever migration path is chosen needs to happen in that 2 months after the spec release.21:26:28
@shadowjonathan:matrix.org[Jonathan]* With “unstable prefixes” there I meant online endpoints, and emitting unstable identifiers, ofc past events are gonna keep those prefixes, I know that21:26:29
@travis:t2l.ioTravisR(or earlier, if practical) 21:27:07
@shadowjonathan:matrix.org[Jonathan]Another thing, btw, can I ask what the idea is behind spec patch releases? Also quarterly?21:27:21
@shadowjonathan:matrix.org[Jonathan]* Another thing, btw, can I ask what the idea is behind spec patch releases? Also quarterly?21:27:28
@travis:t2l.ioTravisRthe spec doesn't have patch releases anymore21:27:41
@travis:t2l.ioTravisRspec releases are planned for quarterly, though21:27:52
@shadowjonathan:matrix.org[Jonathan] It was mentioned in the global version number MSC, what is it called now? 21:28:02
@shadowjonathan:matrix.org[Jonathan]Will it increment with one minor version every quarter?21:28:11
@travis:t2l.ioTravisRThere's just a Major and Minor version. The target is "a spec release every quarter" - if there's nothing to release, so be it. If there's something that demands a Major version increase, we do that.21:28:52
@shadowjonathan:matrix.org[Jonathan]Hmmm, would something like spaces warrant a minor version change?21:30:24
@travis:t2l.ioTravisRyes21:30:32
@travis:t2l.ioTravisRthere's exceptionally few cases where we'd bump the major version21:30:42
@travis:t2l.ioTravisR if a spec release happens, it'll at least be a minor version increase. 21:31:07
@shadowjonathan:matrix.org[Jonathan] I wasn’t asking that, I know that, I was fishing for a question to ask on how a version number would look like if something like spaces wouldn’t warrant a minor version bump 21:31:22
@shadowjonathan:matrix.org[Jonathan] Because then you can’t differentiate between pre-spaces and post-spaces “v1.X”, but alas 21:31:41
@shadowjonathan:matrix.org[Jonathan] * Cuz then you can’t differentiate between pre-spaces and post-spaces “v1.X”, but alas 21:31:53
@travis:t2l.ioTravisR*because21:32:02
@shadowjonathan:matrix.org[Jonathan]Hm? Where?21:32:12
@travis:t2l.ioTravisRwhere you said "cuz" 😛 21:32:24
@shadowjonathan:matrix.org[Jonathan]Oh heck 🤦21:32:34
@shadowjonathan:matrix.org[Jonathan] * Because then you can’t differentiate between pre-spaces and post-spaces “v1.X”, but alas 21:32:48
@travis:t2l.ioTravisRanyways, our rough plan is to just roll spelling mistake fixes into the active release instead of making a dedicated release for them since no one cares about patch versions anyways.21:33:25
@travis:t2l.ioTravisRso if there's anything of value in the spec waiting to be released, it'll be a minor version bump.21:33:47
@travis:t2l.ioTravisR"anything of value" is realistically defined as "not spelling fixes" 21:34:10
@shadowjonathan:matrix.org[Jonathan]I’d wanna sit here all evening asking which ones would and which ones wouldn’t qualify, but I’ll leave it at this, thank you ^^21:34:32

Show newer messages


Back to Room List