Merge pull request #1 from matrix-org/master

Merge from origin
pull/2414/head
iinuwa 5 years ago committed by GitHub
commit 6005c16b60
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23

@ -0,0 +1,2 @@
patreon: matrixdotorg
liberapay: matrixdotorg

@ -0,0 +1,19 @@
---
name: Proposal ready for review
about: A proposal that is ready for review by the core team and community.
title: ''
labels: proposal, proposal-in-review
assignees: ''
---
<!-- Put your "rendered" link here -->
### Pull Request Checklist
<!-- Please read CONTRIBUTING.rst before submitting your pull request -->
* [ ] Pull request includes a [changelog file](https://github.com/matrix-org/matrix-doc/blob/master/CONTRIBUTING.rst#adding-to-the-changelog)
* [ ] Pull request includes a [sign off](https://github.com/matrix-org/matrix-doc/blob/master/CONTRIBUTING.rst#sign-off)
* [ ] Pull request includes ['Rendered' link](https://matrix.org/docs/spec/proposals#process) above.
* [ ] Pull request title and file name include this PR's number as the MSC number.

@ -0,0 +1,16 @@
---
name: Spec clarification/not a proposal
about: A change that's not a spec proposal, such as a clarification to the spec itself.
title: ''
labels: ''
assignees: ''
---
### Pull Request Checklist
<!-- Please read CONTRIBUTING.rst before submitting your pull request -->
* [ ] Pull request includes a [changelog file](https://github.com/matrix-org/matrix-doc/blob/master/CONTRIBUTING.rst#adding-to-the-changelog)
* [ ] Pull request includes a [sign off](https://github.com/matrix-org/matrix-doc/blob/master/CONTRIBUTING.rst#sign-off)
* [ ] Pull request is classified as ['other changes'](https://github.com/matrix-org/matrix-doc/blob/master/CONTRIBUTING.rst#other-changes)

@ -0,0 +1,20 @@
---
name: WIP Proposal
about: A proposal that isn't quite ready for formal review yet.
title: '[WIP] Your Proposal Title'
labels: proposal
assignees: ''
---
<!-- Put your "rendered" link here -->
### Pull Request Checklist
<!-- Please read CONTRIBUTING.rst before submitting your pull request -->
* [ ] Pull request includes a [changelog file](https://github.com/matrix-org/matrix-doc/blob/master/CONTRIBUTING.rst#adding-to-the-changelog)
* [ ] Pull request includes a [sign off](https://github.com/matrix-org/matrix-doc/blob/master/CONTRIBUTING.rst#sign-off)
* [ ] A ['Rendered' link](https://matrix.org/docs/spec/proposals#process) above.
* [ ] Update the title and file name of your proposal to match this PR's number (after opening).
* [ ] Ask in [#matrix-spec:matrix.org](https://matrix.to/#/#matrix-spec:matrix.org) to get this marked as ready for review, once it is ready for review.

@ -31,7 +31,7 @@ paths:
post:
summary: Searches the user directory.
description: |-
Performs a search for users on the homeserver. The homeserver may
Performs a search for users. The homeserver may
determine which subset of users are searched, however the homeserver
MUST at a minimum consider the users the requesting user shares a
room with and those who reside in public rooms (known to the homeserver).

@ -0,0 +1 @@
Minor clarifications to token-based User-Interactive Authentication.

@ -0,0 +1 @@
Minor clarification for what the user directory searches.

@ -93,3 +93,13 @@ of concerns where possible.*
By having a template available, people would know what the desired detail for a proposal is. This is not
considered a risk because it is important that people understand the proposal process from start to end.
## Unstable prefix
*If a proposal is implemented before it is included in the spec, then implementers must ensure that the
implementation is compatible with the final version that lands in the spec. This generally means that
experimental implementations should use `/unstable` endpoints, and use vendor prefixes where necessary.
For more information, see [MSC2324](https://github.com/matrix-org/matrix-doc/pull/2324). This section
should be used to document things such as what endpoints and names are being used while the feature is
in development, the name of the unstable feature flag to use to detect support for the feature, or what
migration steps are needed to switch to newer versions of the proposal.*

@ -216,12 +216,12 @@ response:
}
```
Similarly, the federation endpoints `GET /user/keys/query` and `POST
Similarly, the federation endpoints `POST /user/keys/query` and `POST
/user/devices/{userId}` will include the master and self-signing keys. (It
will not include the user-signing key because it is not intended to be visible
to other users.)
`POST /keys/query`
`POST /user/keys/query`
``` json
{
@ -432,27 +432,31 @@ response:
}
}
},
"master_key": {
"user_id": "@alice:example.com",
"usage": ["master"],
"keys": {
"ed25519:base64+master+public+key": "base64+master+public+key"
},
"signatures": {
"@alice:example.com": {
"ed25519:HIJKLMN": "base64+signature+of+master+key"
"master_keys": {
"@alice:example.com": {
"user_id": "@alice:example.com",
"usage": ["master"],
"keys": {
"ed25519:base64+master+public+key": "base64+master+public+key"
},
"signatures": {
"@alice:example.com": {
"ed25519:HIJKLMN": "base64+signature+of+master+key"
}
}
}
},
"self_signing_key": {
"user_id": "@alice:example.com",
"usage": ["self_signing"],
"keys": {
"ed25519:base64+self+signing+public+key": "base64+self+signing+public+key"
},
"signatures": {
"@alice:example.com": {
"ed25519:base64+master+public+key": "base64+signature"
"self_signing_keys": {
"@alice:example.com": {
"user_id": "@alice:example.com",
"usage": ["self_signing"],
"keys": {
"ed25519:base64+self+signing+public+key": "base64+self+signing+public+key"
},
"signatures": {
"@alice:example.com": {
"ed25519:base64+master+public+key": "base64+signature"
}
}
}
}

@ -0,0 +1,124 @@
# MSC2324: Facilitating early releases of software dependent on spec
*Note*: This is a process change MSC, not a change to the spec itself.
There's currently an unanswered question by the spec process: when is it
safe to start using stable endpoints or to present a feature as "stable"?
Historically this question would receive very different answers depending
on who you asked, so in an effort to come up with a concise answer the
following process change is proposed.
## Proposal
The new process, from start to finish, is proposed as:
1. Have an idea for a feature.
2. Optionally: implement the feature using unstable endpoints, vendor prefixes,
and unstable feature flags as appropriate.
* When using unstable endpoints, they MUST include a vendor prefix. For
example: `/_matrix/client/unstable/com.example/login`. Vendor prefixes
throughout this proposal always use the Java package naming convention.
* Unstable endpoints **do not** inherit from stable (`/r0`) APIs. Previously,
one could access the entirety of the Matrix API through `/unstable` however
this is generally considered a bad practice. Therefore an implementation
can no longer assume that because its feature-specific endpoint exists that
any other endpoint will exist in the same unstable namespace.
* If the client needs to be sure the server supports the feature, an unstable
feature flag that MUST be vendor prefixed is to be used. This kind of flag
shows up in the `unstable_features` field of `/versions` as, for example,
`com.example.new_login`.
* You can ship the feature at *any* time, so long as you are able to accept
the technical debt that results from needing to provide adequate backwards
and forwards compatibility for the vendor prefixed implementation. The
implementation MUST support the flag disappearing and be generally safe for
users. Note that implementations early in the MSC review process may also be
required to provide backwards compatibility with earlier editions of the
proposal.
* If you don't want to support the technical debt (or if it's impossible to
provide adequate backwards/forwards compatibility - e.g. a user authentication
change which can't be safely rolled back), do not implement the feature and
wait for Step 7.
* If at any point the idea changes, the feature flag should also change so
that implementations can adapt as needed.
3. In parallel, or ahead of implementation, open an MSC and solicit review.
4. Before a FCP (Final Comment Period) can be called, the Spec Core Team will
require that evidence to prove the MSC works be presented. A typical example
of this is an implementation of the MSC (which does not necessarily need to have been shipped anywhere).
5. FCP is gone through, and assuming nothing is flagged the MSC lands.
6. A spec PR is written to incorporate the changes into Matrix.
7. A spec release happens.
8. Implementations switch to using stable prefixes (e.g.: `/r0`) if the server
supports the specification version released. If the server doesn't advertise
the specification version, but does have the feature flag, unstable prefixes
should still be used.
9. A transition period of about 2 months starts immediately after the spec release, before
implementations start to loudly encourage other implementations to switch to stable
endpoints. For example, the Synapse team should start asking the Riot team to
support the stable endpoints (as per Step 8) 2 months after the spec release if they
haven't already.
It's worth repeating that this process generally only applies if the implementation
wants to ship the feature ahead of the spec being available. By doing so, it takes
on the risk that the spec/MSC may change and it must adapt. If the implementation
is unable to take on that risk, or simply doesn't mind waiting, it should go through
the spec process without shipping an unstable implementation.
To help MSCs get incorporated by implementations as stable features, the spec core
team plans to release the specification more often. How often is undefined and is
largely a case-by-case basis.
To reiterate:
* Implementations MUST NOT use stable endpoints before the MSC is in the spec. This
includes NOT using stable endpoints post-FCP.
* Implementations CAN ship features that are exposed by default to users before an
MSC has been merged to the spec, provided they follow the process above.
* Implementations SHOULD be wary of the technical debt they are incurring by moving
faster than the spec.
To clarify:
* The vendor prefix is chosen by the developer of the feature, using the Java package
naming convention. For example, `org.matrix` is the foundation's vendor prefix.
* The vendor prefixes, unstable feature flags, and unstable endpoints should be included
in the MSC so other developers can benefit. The MSC MUST still say what the stable
endpoints are to look like.
### Specific examples outside of the client-server API
There are some instances where a change might be made outside of the client-server API,
which is where much of this proposal is targetted. The general spirit of the process
should be followed where possible, if implementations decide to work ahead of spec releases.
#### Room versions
When a new room version is needed, implementations MUST use vendor-prefixed versions
before using the namespace reserved for Matrix (see https://matrix.org/docs/spec/#room-versions).
A room version is considered released once it is listed as an "available room version" in
the spec. Often a new room version is accompanied with a server-server API release, but
doesn't have to be.
#### Server-server / Identity / Push / Appservice API
These APIs don't yet have a `/versions` endpoint or similar. Typically behaviour changes in
these APIs are introduced with backwards compatibility in mind (try X and if that fails fall
back to Y) and therefore don't always need a flag to indicate support. If a flag were to
be required, an `unstable_features` or similar array would need to be exposed somewhere.
#### Changes to request/response parameters
Parameters being added to request/response bodies and query strings MUST be vendor-prefixed
per the proposed process. For example, a new JSON field might be `{"org.matrix.example": true}`
with the proposal being for `example` being added. A query string parameter would be prefixed
the same way: `?org.matrix.example=true`.
If the MSC is simply adding fields to already-versioned endpoints, it does not need to put
the whole endpoint into the `/unstable` namespace provided the new parameters are prefixed
appropriately.
#### .well-known and other APIs that can't be versioned
Best effort is appreciated. Typically these endpoints will be receiving minor behavioural
changes or new fields. New fields should be appropriately prefixed, and behaviour changes
should be rolled out cautiously by implementations (waiting until after FCP has concluded
is probably best to ensure there's no major problems with the new behaviour).

@ -0,0 +1,81 @@
# Allowing Reasons in all Membership Events
Currently users can specify a reason for kicking or banning users in a room
that both the target and other users in a room can see. This is useful to
explain *why* an action without having to send separate messages.
The proposal extends this to all events, including invites, invite rejections
and leaves for similar reasons.
## Proposal
Allow `reason` field to be specified for all of the following APIs:
```
POST /_matrix/client/r0/rooms/{roomId}/invite
POST /_matrix/client/r0/rooms/{roomId}/leave
POST /_matrix/client/r0/rooms/{roomId}/kick
POST /_matrix/client/r0/rooms/{roomId}/ban
POST /_matrix/client/r0/rooms/{roomId}/unban
POST /_matrix/client/r0/rooms/{roomId}/join
POST /_matrix/client/r0/join/{roomIdOrAlias}
PUT /_matrix/client/r0/rooms/{roomId}/state/m.room.member/{userID}
```
If specified the `reason` field will be added to the generated membership
event's content.
*Note: `/state/m.room.member` API currently allows this as clients can specify
arbitrary content already*
Clients may choose to display the reason for membership events in a room,
though may not do so if e.g. they have collapsed a set of membership changes.
Clients should not display an invite reason by default to the invitee as this
allows a classic abuse and harassment vector. However, clients may wish to show
invite reasons from known¹ senders, or have a button that allows viewing any
invite reason.
## Use Cases
Some basic use cases and examples are given below.
### Kick/ban
Kicking and banning already allow specifying a reason to allow giving a reason
for the moderation action (e.g. "Banned for spamming").
### Leaving a Room
A user may wish to leave a room e.g. because the room is no longer relevant
to them, allowing them to specify a reason means they can easily step out of
the room quietly without having to send a message to explain their actions.
### Invite
This can be useful to give some context for an invite, e.g. "Alice invites Bob
to get some feedback on the membership reasons MSC".
### Rejecting an Invite
If Alice has invited Bob (and many others) to a room to discuss going to a
concert then Bob may wish to simply reject the invite if he can't make it.
Adding a "will be out of town" reason to the rejection helps Alice to know why
her invite was rejected.
### Joining room
Adding a reason for joining could be used e.g. by automated bots to say why
they're joining. For example a bridge bot may join a room when asked to bridge
the room to an external network, in which case they may have a reason such as
"BridgeBot joined to bridge the room to RemoteNetwork at the request of Alice".
## Potential Issues
The main issue here is ensuring that the invite reason cannot be used as an
abuse vector, however if clients follow the recommendations above this concern
should be mitigated.
---
¹ This is left up to implementations to decide, if they wish to do so.

@ -740,22 +740,23 @@ To use this authentication type, clients should submit an auth dict as follows:
"session": "<session ID>"
}
The ``nonce`` should be a random string generated by the client for the
request. The same ``nonce`` should be used if retrying the request.
A client may receive a login ``token`` via some external service, such as email
or SMS. Note that a login token is separate from an access token, the latter
providing general authentication to various API endpoints.
The ``txn_id`` may be used by the server to disallow other devices from using
the token, thus providing "single use" tokens while still allowing the device
to retry the request. This would be done by tying the token to the ``txn_id``
Additionally, the server must encode the user ID in the ``token``; there is
therefore no need for the client to submit a separate username.
The ``txn_id`` should be a random string generated by the client for the
request. The same ``txn_id`` should be used if retrying the request. The
``txn_id`` may be used by the server to disallow other devices from using the
token, thus providing "single use" tokens while still allowing the device to
retry the request. This would be done by tying the token to the ``txn_id``
server side, as well as potentially invalidating the token completely once the
device has successfully logged in (e.g. when we receive a request from the
newly provisioned access_token).
The server must encode the user id in the ``token``. There is therefore no need
for the client to submit a separate username.
OAuth2-based
<<<<<<<<<<<<

@ -94,7 +94,7 @@ results of the resolution so far.
passes authentication in :math:`R` and add it to :math:`R`.
A *conflict* occurs between states where those states have different
``event_ids`` for the same ``(state_type, state_key)``. The events thus
``event_ids`` for the same ``(event_type, state_key)``. The events thus
affected are said to be *conflicting* events.

Loading…
Cancel
Save