-
Notifications
You must be signed in to change notification settings - Fork 434
MSC3916: Authentication for media #3916
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Changes from 17 commits
7606e53
3076de0
55303b5
d601637
c2ae25e
8351ebe
106ce55
a14c4af
92eba2c
a76d97f
8bb5159
e1e8a6a
71b8db4
1c864a3
e5c9316
41d2aa2
d73025b
656dfb8
87c08e0
1fe8d71
aac1909
0b4f2c9
f1777c2
4fe23e5
db90b92
File filter
Filter by extension
Conversations
Jump to
Diff view
Diff view
There are no files selected for viewing
|
Member
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Reviewers: Please compare this MSC against #2461 , which is proposed for rejection/closure.
Member
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. @turt2live I think we can resolve this since #2461 is closed?
Member
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. it's still useful to compare the MSCs, given the ideas and concepts of 2461 are meant to be incorporated here. |
| Original file line number | Diff line number | Diff line change | ||||
|---|---|---|---|---|---|---|
| @@ -0,0 +1,367 @@ | ||||||
| # MSC3916: Authentication for media access, and new endpoint names | ||||||
|
|
||||||
| Currently, access to media in Matrix has a number of problems including the following: | ||||||
|
|
||||||
| * The only protection for media is the obscurity of the URL, and URLs are | ||||||
| easily leaked (eg accidental sharing, access | ||||||
| logs). [synapse#2150](https://github.com/matrix-org/synapse/issues/2150) | ||||||
| * Anybody (including non-matrix users) can cause a homeserver to copy media | ||||||
| into its local | ||||||
| store. [synapse#2133](https://github.com/matrix-org/synapse/issues/2133) | ||||||
| * When a media event is redacted, the media it used remains visible to all. | ||||||
| [synapse#1263](https://github.com/matrix-org/synapse/issues/1263) | ||||||
| * There is currently no way to delete | ||||||
| media. [matrix-spec#226](https://github.com/matrix-org/matrix-spec/issues/226) | ||||||
| * If a user requests GDPR erasure, their media remains visible to all. | ||||||
| * When all users leave a room, their media is not deleted from the server. | ||||||
|
|
||||||
| These problems are all difficult to address currently, because access to media | ||||||
| is entirely unauthenticated. The first step for a solution is to require user | ||||||
| authentication. Once that is done, it will be possible to impose authorization | ||||||
| requirements to address the problems mentioned above. (See, for example, | ||||||
| [MSC3911](https://github.com/matrix-org/matrix-spec-proposals/pull/3911) which | ||||||
| builds on top of this MSC.) | ||||||
|
|
||||||
| This proposal supersedes [MSC1902](https://github.com/matrix-org/matrix-spec-proposals/pull/1902). | ||||||
|
|
||||||
| ## Proposal | ||||||
|
|
||||||
| 1. New endpoints | ||||||
|
|
||||||
| The existing `/_matrix/media/v3/` endpoints become deprecated, and new | ||||||
| endpoints under the `/_matrix/client` and `/_matrix/federation` | ||||||
| hierarchies are introduced. Removal of the deprecated endpoints would be a | ||||||
| later MSC under [the deprecation policy](https://spec.matrix.org/v1.10/#deprecation-policy). | ||||||
|
|
||||||
| The following table below shows a mapping between deprecated and new endpoint: | ||||||
|
|
||||||
| | Deprecated | Client-Server | Federation | | ||||||
| | ---------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------- | ------------------------------------------------------------------- | | ||||||
| | [`GET /_matrix/media/v3/preview_url`](https://spec.matrix.org/v1.6/client-server-api/#get_matrixmediav3preview_url) | `GET /_matrix/client/v1/media/preview_url` | - | | ||||||
| | [`GET /_matrix/media/v3/config`](https://spec.matrix.org/v1.6/client-server-api/#get_matrixmediav3config) | `GET /_matrix/client/v1/media/config` | - | | ||||||
| | [`GET /_matrix/media/v3/download/{serverName}/{mediaId}`](https://spec.matrix.org/v1.6/client-server-api/#get_matrixmediav3downloadservernamemediaid) | `GET /_matrix/client/v1/media/download/{serverName}/{mediaId}` | `GET /_matrix/federation/v1/media/download/{mediaId}` | | ||||||
| | [`GET /_matrix/media/v3/download/{serverName}/{mediaId}/{fileName}`](https://spec.matrix.org/v1.6/client-server-api/#get_matrixmediav3downloadservernamemediaidfilename) | `GET /_matrix/client/v1/media/download/{serverName}/{mediaId}/{fileName}` | - | | ||||||
| | [`GET /_matrix/media/v3/thumbnail/{serverName}/{mediaId}`](https://spec.matrix.org/v1.6/client-server-api/#get_matrixmediav3thumbnailservernamemediaid) | `GET /_matrix/client/v1/media/thumbnail/{serverName}/{mediaId}` | - | | ||||||
|
turt2live marked this conversation as resolved.
Outdated
|
||||||
|
|
||||||
| **Note**: [`POST /_matrix/media/v3/upload`](https://spec.matrix.org/v1.6/client-server-api/#post_matrixmediav3upload) | ||||||
| and [`GET /_matrix/media/v1/create`](https://spec.matrix.org/v1.10/client-server-api/#post_matrixmediav1create) | ||||||
|
turt2live marked this conversation as resolved.
Outdated
|
||||||
| are **not** modified or deprecated by this MSC: it is intended that they be brought into line with the other | ||||||
| endpoints by a future MSC, such as [MSC3911](https://github.com/matrix-org/matrix-spec-proposals/pull/3911). | ||||||
|
|
||||||
| **Note**: `/thumbnail` does not have a federation endpoint. It appears as though | ||||||
| no servers request thumbnails over federation, and so it is not supported here. | ||||||
| A later MSC may introduce such an endpoint. | ||||||
|
turt2live marked this conversation as resolved.
Outdated
|
||||||
|
|
||||||
| The new `/download` and `/thumbnail` endpoints additionally drop the `?allow_redirect` | ||||||
| query parameters. Instead, the endpoints behave as though `allow_redirect=true` was | ||||||
| set, regardless of actual value. See [this comment on MSC3860](https://github.com/matrix-org/matrix-spec-proposals/pull/3860/files#r1005176480) | ||||||
| for details. | ||||||
|
turt2live marked this conversation as resolved.
Outdated
|
||||||
|
|
||||||
| After this proposal is released in a stable version of the specification, servers | ||||||
| which support the new `download` and `thumbnail` endpoints SHOULD cease to serve | ||||||
| newly uploaded media from the unauthenticated versions. This includes media | ||||||
|
turt2live marked this conversation as resolved.
Outdated
|
||||||
| uploaded by local users and requests for not-yet-cached remote media. This is | ||||||
| done with a 404 `M_NOT_FOUND` error, as though the media doesn't exist. Servers | ||||||
| SHOULD consider their local ecosystem impact before freezing the endpoints. For | ||||||
| example, ensuring that common bridges and clients will continue to work, and | ||||||
| encouraging updates to those affected projects as needed. | ||||||
|
|
||||||
| 2. Removal of `allow_remote` parameter from `/download` | ||||||
|
|
||||||
| The current | ||||||
| [`/download`](https://spec.matrix.org/v1.6/client-server-api/#get_matrixmediav3downloadservernamemediaid) | ||||||
| and | ||||||
| [`/thumbnail`](https://spec.matrix.org/v1.6/client-server-api/#get_matrixmediav3thumbnailservernamemediaid) | ||||||
| endpoints take an `allow_remote` query parameter, indicating whether the | ||||||
| server should request remote media from other servers. This is redundant | ||||||
| with the new endpoints, so will not be supported. | ||||||
|
|
||||||
| Servers MUST NOT return remote media from `GET /_matrix/federation/v1/media/download`. The | ||||||
| `serverName` is omitted from the endpoint's path to strongly enforce this - the `mediaId` in | ||||||
| a request is assumed to be scoped to the target server. | ||||||
|
|
||||||
| `/_matrix/client/v1/media/download` and | ||||||
| `/_matrix/client/v1/media/thumbnail` return remote media as normal. | ||||||
|
|
||||||
| 3. Authentication on all endpoints | ||||||
|
|
||||||
| Currently, the `/download` and `/thumbnail` endpoints have no authentication | ||||||
| requirements. Under this proposal, the new endpoints will be authenticated | ||||||
| the same way as other endpoints: they will require an `Authorization` header | ||||||
| which must be `Bearer {accessToken}` for `/_matrix/client`, or the signature | ||||||
|
turt2live marked this conversation as resolved.
|
||||||
| for `/_matrix/federation`. | ||||||
|
turt2live marked this conversation as resolved.
turt2live marked this conversation as resolved.
|
||||||
|
|
||||||
| The query string `?access_token` approach is not supported on the new endpoints, | ||||||
| as it is [deprecated](https://github.com/matrix-org/matrix-spec-proposals/blob/main/proposals/4126-deprecate-query-string-auth.md) | ||||||
|
Member
Author
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. it's easier for a reader to go from the PR to the doc than the other way around, so could we do:
Suggested change
Member
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. The PR doesn't show any changes to accepted MSCs, unfortunately. Ideally this would be a link to the spec, but unstable has the same longevity problems. |
||||||
| and [pending removal](https://github.com/matrix-org/matrix-spec-proposals/pull/4127). | ||||||
| See those MSCs for details. | ||||||
|
turt2live marked this conversation as resolved.
Outdated
|
||||||
|
|
||||||
| **Note**: This fixes [matrix-spec#313](https://github.com/matrix-org/matrix-spec/issues/313). | ||||||
|
|
||||||
| 4. Updated response format | ||||||
|
|
||||||
| * For the new `/_matrix/client` endpoints, the response format is the same as | ||||||
| the corresponding original endpoints. | ||||||
|
|
||||||
| * To enable future expansion, for the new `/_matrix/federation` endpoints, | ||||||
| the response is | ||||||
| [`multipart/mixed`](https://www.w3.org/Protocols/rfc1341/7_2_Multipart.html) | ||||||
|
turt2live marked this conversation as resolved.
|
||||||
| content with exactly two parts: the first MUST be a JSON object (and should have a | ||||||
| `Content-type: application/json` header), and the second MUST be the media item | ||||||
| as per the original endpoints. | ||||||
|
turt2live marked this conversation as resolved.
Outdated
|
||||||
|
|
||||||
| No properties are yet specified for the JSON object to be returned. One | ||||||
| possible use is described by [MSC3911](https://github.com/matrix-org/matrix-spec-proposals/pull/3911). | ||||||
|
|
||||||
| An example response: | ||||||
|
|
||||||
| ``` | ||||||
| Content-Type: multipart/mixed; boundary=gc0p4Jq0M2Yt08jU534c0p | ||||||
|
|
||||||
| --gc0p4Jq0M2Yt08jU534c0p | ||||||
| Content-Type: application/json | ||||||
|
|
||||||
| {} | ||||||
|
|
||||||
| --gc0p4Jq0M2Yt08jU534c0p | ||||||
| Content-Type: text/plain | ||||||
|
|
||||||
| This media is plain text. Maybe somebody used it as a paste bin. | ||||||
|
|
||||||
| --gc0p4Jq0M2Yt08jU534c0p | ||||||
| ``` | ||||||
|
turt2live marked this conversation as resolved.
|
||||||
|
|
||||||
| The second part (media item bytes) MAY include a [`Location` header](https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Location) | ||||||
| to point to the raw media object instead of having bytes itself. Servers | ||||||
| SHOULD NOT cache the `Location` header's value as the responding server may | ||||||
| have applied time limits on its validity. Servers which don't immediately | ||||||
| download the media from the provided URL should re-request the media and | ||||||
| metadata from the `/download` endpoint when ready for the media bytes. | ||||||
|
turt2live marked this conversation as resolved.
|
||||||
|
|
||||||
| The `Location` header's URL does *not* require authentication, as it will | ||||||
| typically be served by a CDN or other non-matrix server (thus being unable | ||||||
| to verify any `X-Matrix` signatures, for example). | ||||||
|
|
||||||
| An example response with a `Location` redirect would be: | ||||||
|
|
||||||
| ``` | ||||||
| Content-Type: multipart/mixed; boundary=gc0p4Jq0M2Yt08jU534c0p | ||||||
|
|
||||||
| --gc0p4Jq0M2Yt08jU534c0p | ||||||
| Content-Type: application/json | ||||||
|
|
||||||
| {} | ||||||
|
|
||||||
| --gc0p4Jq0M2Yt08jU534c0p | ||||||
| Content-Type: text/plain | ||||||
|
turt2live marked this conversation as resolved.
Outdated
|
||||||
| Location: https://cdn.example.org/ab/c1/2345.txt | ||||||
|
tulir marked this conversation as resolved.
|
||||||
|
|
||||||
| --gc0p4Jq0M2Yt08jU534c0p | ||||||
| ``` | ||||||
|
|
||||||
| If the server were to `curl https://cdn.example.org/ab/c1/2345.txt`, it'd | ||||||
| get: | ||||||
|
|
||||||
| ``` | ||||||
| This media is plain text. Maybe somebody used it as a paste bin. | ||||||
| ``` | ||||||
|
|
||||||
| 5. Backwards compatibility mechanisms | ||||||
|
|
||||||
| a. Backwards compatibility with older servers: if a client or requesting server | ||||||
| receives a 404 error with `M_UNRECOGNIZED` error code in response to a request | ||||||
| using the new endpoints, they may retry the request using the deprecated | ||||||
| endpoint. Servers and clients should note the [`M_UNRECOGNIZED`](https://spec.matrix.org/v1.10/client-server-api/#common-error-codes) | ||||||
| error code semantics. | ||||||
|
|
||||||
| b. Backwards compatibility with older clients and federating servers: mentioned | ||||||
| in Part 1 of this proposal, servers *may* freeze unauthenticated media access | ||||||
| once stable authenticated endpoints are available. This may lead to client and | ||||||
| server errors for new media. Both clients and servers are strongly encouraged | ||||||
| to update as soon as possible, before servers freeze unauthenticated media | ||||||
| access. | ||||||
|
|
||||||
| 6. Removal of `allow_redirect` parameter from `/download` and `/thumbnail` | ||||||
|
|
||||||
| Clients MUST expect a 307 or 308 redirect when calling the new `/download` | ||||||
| and `/thumbnail` Client-Server API endpoints. | ||||||
|
|
||||||
| Servers MUST expect the `Location` header in the media part of the new Server-Server | ||||||
| API `/download` endpoint. Servers MUST NOT respond with a 307 or 308 redirect at | ||||||
| the top level for the endpoint - they can only redirect within the media part | ||||||
| itself. | ||||||
|
|
||||||
| ### Effects on client applications | ||||||
|
|
||||||
| Naturally, implementations will be required to provide `Authorization` headers | ||||||
| when accessing the new endpoints. This will be simple in some cases, but rather | ||||||
| more involved in others. This section considers some of those cases. | ||||||
|
|
||||||
| #### IRC/XMPP bridges | ||||||
|
|
||||||
| Possibly the largest impact will be on IRC and XMPP bridges. Since IRC and | ||||||
| XMPP have no media repository of their own, these bridges currently transform | ||||||
| `mxc:` URIs into `https://<server>/_matrix/media/v3/download/` URIs and forward | ||||||
| those links to the remote platform. This will no longer be a viable option. | ||||||
|
|
||||||
| One potential solution is for the bridges to provide a proxy. | ||||||
|
|
||||||
| In this scenario, the bridge would have a secret HMAC key. When it | ||||||
| receives a matrix event referencing a piece of media, it should create a new URI | ||||||
| referencing the media, include an HMAC to prevent tampering. For example: | ||||||
|
|
||||||
| ``` | ||||||
| https://<bridge_server>/media/{originServerName}/{mediaId}?mac={hmac} | ||||||
| ``` | ||||||
|
|
||||||
| When the bridge later receives a request to that URI, it checks the hmac, | ||||||
| and proxies the request to the homeserver, using its AS access | ||||||
| token in the `Authorization` header. | ||||||
|
|
||||||
| The bridge might also choose to embed information such as the room that | ||||||
| referenced the media, and the time that the link was generated, in the URL. | ||||||
| Such mechanisms would allow the bridge to impose controls such as: | ||||||
|
|
||||||
| * Limiting the time a media link is valid for. Doing so would help prevent | ||||||
| visibility to users who weren't participating in the chat. | ||||||
|
|
||||||
| * Rate-limiting the amount of media being shared in a particular room (in other | ||||||
| words, avoiding the use of Matrix as a Warez distribution system). | ||||||
|
|
||||||
| #### Icons for "social login" flows | ||||||
|
|
||||||
| When a server supports multiple login providers, it provides the client with | ||||||
| icons for the login providers as `mxc:` media URIs. These must be accessible | ||||||
| without authentication (because the client has no access token at the time the | ||||||
| icons are displayed). | ||||||
|
|
||||||
| This remains a somewhat unsolved problem. Possibly the clients can continue | ||||||
|
turt2live marked this conversation as resolved.
Outdated
|
||||||
| to call the legacy `/_matrix/media/v3/download` URI for now: ultimately this | ||||||
| problem will be solved by the transition to OIDC. Alternatively, a dedicated | ||||||
| API could be provided or permission to use HTTP(S) URLs to access the icons. | ||||||
| This support would come from a different MSC. | ||||||
|
|
||||||
| (This was previously discussed in | ||||||
| [MSC2858](https://github.com/matrix-org/matrix-spec-proposals/pull/2858#discussion_r543513811).) | ||||||
|
|
||||||
| ## Potential issues | ||||||
|
turt2live marked this conversation as resolved.
turt2live marked this conversation as resolved.
turt2live marked this conversation as resolved.
|
||||||
|
|
||||||
| * Setting the `Authorization` header is particularly annoying for web clients. | ||||||
| Service workers are seemingly the best option, though other options include | ||||||
| locally-cached `blob:` URIs. Clients should note that caching media can lead | ||||||
| to significant memory usage, particularly for large media. Service workers by | ||||||
| comparison allow for proxy-like behaviour. | ||||||
|
turt2live marked this conversation as resolved.
|
||||||
|
|
||||||
| Cookies are a plausible mechanism for sharing session information between | ||||||
| requests without having to set headers, though would be a relatively bespoke | ||||||
| authentication method for Matrix. Additionally, many Matrix users have cookies | ||||||
| disabled due to the advertising and tracking use cases common across the web. | ||||||
|
Comment on lines
+289
to
+290
Member
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Is this really true? What data do you have to back this up? It seems like a fairly elegant solution.
Member
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Not hard data, but for the last 6 years it's felt like every third user for Element Web has disabled these things and more :p
Member
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Mmm, the latter seems spurious since cookies and localstorage will almost always be enabled/disabled as a pair, and Element has been relying on localstorage for years. The former is true in that setting a cookie for the homeserver would be a little weird. In my mind, the pros / cons are:
I'm not going to make this a concern as I don't think it's blocking, but it might be worth a quick sanity check to make sure we really are committing to the right thing. |
||||||
|
|
||||||
| * Users will be unable to copy links to media from web clients to share out of | ||||||
| band. This is considered a feature, not a bug. | ||||||
|
|
||||||
| * Over federation, the use of the `Range` request header on `/download` becomes | ||||||
| unclear as it could affect either or both parts of the response. There does not | ||||||
| appear to be formal guidance in [RFC 9110](https://www.rfc-editor.org/rfc/rfc9110#field.range) | ||||||
| either. There are arguments for affecting both and either part equally. Typically, | ||||||
| such a header would be used to resume failed downloads, though servers are | ||||||
| already likely to discard received data and fail the associated client requests | ||||||
| when the federation request fails. Therefore, servers are unlikely to use `Range` | ||||||
| at all. As such, this proposal does not make a determination on how `Range` | ||||||
| should be handled, and leaves it as an HTTP specification interpretation problem | ||||||
| instead. | ||||||
|
|
||||||
| * The `Location` header support on the new `/download` endpoint could add a bit | ||||||
| of complexity to servers, though given the alternative of supporting CDNs and | ||||||
| similar is to place complexity into "edge workers" to mutate the response value. | ||||||
| Though the Matrix spec would be "simpler", the edge worker setup would be | ||||||
| fragmented where we have an opportunity for a common standard. | ||||||
|
|
||||||
|
turt2live marked this conversation as resolved.
|
||||||
| ## Alternatives | ||||||
|
turt2live marked this conversation as resolved.
|
||||||
|
|
||||||
| * Allow clients to upload media which does not require authentication (for | ||||||
|
turt2live marked this conversation as resolved.
|
||||||
| example via a `public=true` query parameter). This might be particularly | ||||||
| useful for IRC/XMPP bridges, which could upload any media they encounter to | ||||||
| the homeserver's repository. | ||||||
|
|
||||||
| The danger with this is that is that there's little stopping clients | ||||||
| continuing to upload media as "public", negating all of the benefits in this | ||||||
| MSC. It might be ok if media upload it was restricted to certain privileged | ||||||
| users, or applied after the fact by a server administrator. | ||||||
|
|
||||||
| * We could simply require that `Authorization` headers be given when calling | ||||||
| the existing endpoints. However, doing so would make it harder to evaluate | ||||||
| the proportion of clients which have been updated, and it is a good | ||||||
| opportunity to bring these endpoints into line with the rest of the | ||||||
| client-server and federation APIs. | ||||||
|
|
||||||
| * There's no real need to rename `GET /_matrix/media/v3/preview_url` and `GET | ||||||
| /_matrix/media/v3/config` at present, and we could just leave them in | ||||||
| place. However, changing them at the same time makes the API more consistent. | ||||||
|
|
||||||
| Conversely, we should make sure to rename `POST | ||||||
| /_matrix/media/v3/upload` and `GET /_matrix/media/v3/create`. The reason to | ||||||
| delay doing so is because MSC3911 will make more substantial changes to these | ||||||
| endpoints, requiring another rename, and it is expected that both proposals | ||||||
| will be mergeed near to the same time as each other (so a double rename will | ||||||
|
turt2live marked this conversation as resolved.
Outdated
|
||||||
| be confusing and unnecessary). However, if MSC3911 is delayed or rejected, we | ||||||
| should reconsider this. | ||||||
|
Comment on lines
+339
to
+340
Member
Author
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. We are still planning to fast-follow with MSC3911, right?
Member
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Hopefully, yes. It won't be in the same spec release though. |
||||||
|
|
||||||
| * Rather than messing with multipart content, have a separate endpoint for | ||||||
| servers to get the metadata for a media item. That would mean two requests, | ||||||
| but might make more sense than `/download` providing the info directly. | ||||||
|
turt2live marked this conversation as resolved.
Outdated
|
||||||
|
|
||||||
| This is a plausible approach with no significant upsides or downsides when | ||||||
| compared to multipart responses. | ||||||
|
|
||||||
| Similarly, custom headers could be used to carry the metadata on the response, | ||||||
| though again, there are no significant upsides or downsides to doing so. | ||||||
|
|
||||||
| Readers may wish to refer to [this thread](https://github.com/matrix-org/matrix-spec-proposals/pull/3916/files#r1586878787) | ||||||
| on the MSC which covers the majority of the pros and cons for all 3 approaches. | ||||||
|
|
||||||
| ### Compared to MSC3796 (MSC701) | ||||||
|
|
||||||
| [MSC701/3796](https://github.com/matrix-org/matrix-spec-proposals/issues/3796) | ||||||
| introduces a concept of "content tokens" which have authentication tie-in to | ||||||
| prevent anonymous users from accessing media. This is a similar problem space | ||||||
| to this proposal, though deals more in the event-to-media linking space instead. | ||||||
| Although the MSC is an early sketch, it's unclear if the problems noted on the | ||||||
| MSC itself are feasibly resolvable. | ||||||
|
|
||||||
| ### Compared to MSC2461 | ||||||
|
|
||||||
| [MSC2461](https://github.com/matrix-org/matrix-spec-proposals/pull/2461) adds | ||||||
| authentication to the existing media endpoints, which as noted here in the | ||||||
| Alternatives is not likely to roll out quickly and leaves an inconsistency in | ||||||
| the spec. MSC2461 also introduces a client-visible flag for which kinds of media | ||||||
| may require authentication, making it similar to the alternative listed above | ||||||
| where on the federation side we could have two endpoints: one for information | ||||||
| and one for the media itself. MSC2461 simply makes the information client-visible | ||||||
| instead of server-visible. | ||||||
|
|
||||||
| ## Unstable prefix | ||||||
|
|
||||||
| While this proposal is in development, the new endpoints should be named as follows: | ||||||
|
|
||||||
| * `GET /_matrix/client/unstable/org.matrix.msc3916/media/preview_url` | ||||||
| * `GET /_matrix/client/unstable/org.matrix.msc3916/media/config` | ||||||
| * `GET /_matrix/client/unstable/org.matrix.msc3916/media/download/{serverName}/{mediaId}` | ||||||
| * `GET /_matrix/client/unstable/org.matrix.msc3916/media/download/{serverName}/{mediaId}/{fileName}` | ||||||
| * `GET /_matrix/client/unstable/org.matrix.msc3916/media/thumbnail/{serverName}/{mediaId}` | ||||||
| * `GET /_matrix/federation/unstable/org.matrix.msc3916.v2/media/download/{mediaId}` | ||||||
| * **Note**: This endpoint has a `.v2` in its unstable identifier due to the MSC changing after | ||||||
| initial implementation. The original unstable endpoint has a `serverName` and may still be | ||||||
| supported by some servers: `GET /_matrix/federation/unstable/org.matrix.msc3916/media/download/{serverName}/{mediaId}` | ||||||
|
|
||||||
| The `serverName` was later dropped in favour of explicit scoping. See `allow_remote` details | ||||||
| in the MSC body for details. | ||||||
|
|
||||||
| In a prior version of this proposal, the federation API included a thumbnail endpoint. | ||||||
| It was removed due to lack of perceived usage. Servers which implemented the unstable | ||||||
| version will have done so under `GET /_matrix/federation/unstable/org.matrix.msc3916/media/thumbnail/{serverName}/{mediaId}`. | ||||||
| The client-server thumbnail endpoint is unaffected by this change. | ||||||
|
|
||||||
|
turt2live marked this conversation as resolved.
|
||||||
| ## Dependencies | ||||||
|
|
||||||
| None. | ||||||
Uh oh!
There was an error while loading. Please reload this page.