Fix https://github.com/matrix-org/matrix-public-archive/issues/80
```
RethrownError: Unable to fetch rooms from room directory (homeserver=http://localhost:8008/)
searchTerm=, paginationToken=undefined, limit=9
at matrix-public-archive\server\routes\room-directory-routes.js:55:13
--- Original Error ---
Error: HTTP Error Response: 500 Internal Server Error: {"errcode":"M_UNKNOWN","error":"Internal server error"}
URL=http://localhost:8008/_matrix/client/v3/publicRooms?
at checkResponseStatus (matrix-public-archive\server\lib\fetch-endpoint.js:21:11)
at processTicksAndRejections (node:internal/process/task_queues:96:5)
at async fetchEndpoint (matrix-public-archive\server\lib\fetch-endpoint.js:38:3)
at async fetchEndpointAsJson (matrix-public-archive\server\lib\fetch-endpoint.js:63:15)
at async fetchPublicRooms (matrix-public-archive\server\lib\matrix-utils\fetch-public-rooms.js:26:26)
at async matrix-public-archive\server\tracing\trace-utilities.js:31:24
at async matrix-public-archive\server\routes\room-directory-routes.js:45:62
```
Regressed in https://github.com/matrix-org/matrix-public-archive/pull/61 where we tried to serve this under `/css/hydrogen-styles.css` but it doesn't work because all of the image and font references in the CSS file expect it to be at the domain root so just reverted back to serving at the root `/`.
This isn't spawning from any previous security issue. Just adding an extra check to help ensure we don't ever regress this in the future.
```
AssertionError [ERR_ASSERTION]: We should not be leaking the `config.matrixAccessToken` to the Hydrogen render function because this will reach the client!
at renderHydrogenToString (matrix-public-archive\server\hydrogen-render\render-hydrogen-to-string.js:24:3)
at renderHydrogenVmRenderScriptToPageHtml (matrix-public-archive\server\hydrogen-render\render-hydrogen-vm-render-script-to-page-html.js:22:36)
at matrix-public-archive\server\routes\room-directory-routes.js:53:28
at processTicksAndRejections (node:internal/process/task_queues:96:5)
```
1. Add surrounding messages to the given messages so we have a full screen of content to make it feel lively even in quiet rooms
- As you scroll around the timeline across different days, the date changes in the URL, calendar, etc
2. Add summary item to the bottom of the timeline that explains if we couldn't find any messages in the specific day requested
- Also allows you to the jump to the next activity in the room. Adds `/:roomId/jump?ts=xxx&dir=[f|b]` to facilitate this.
- Part of https://github.com/matrix-org/matrix-public-archive/issues/46
3. Add developer options modal which is linked from the bottom of the right-panel
- Adds an option so you can debug the `IntersectionObserver` and how it's selecting the active day from the top-edge of the scroll viewport.
- In the future, this will also include a nice little visualization of the backend timing traces
Bigger more clickable buttons on mobile. Feels a lot better on a phone. The buttons already have invisible margin in their hitbox but the bigger size makes your thumb less cramped to the edge to click them.
Also reduce the container padding so it feels more balanced in the single column card layout.
Follow-up to https://github.com/matrix-org/matrix-public-archive/pull/36
Render pipeline separation of concerns:
1. Run in `child_process`
2. Hydrogen render
It's now just a generic `child_process` runner that runs the Hydrogen render in it. This eliminates the windy path of the 1-4 steps that was only held together by the file names themselves.
Follow-up to https://github.com/matrix-org/matrix-public-archive/pull/51
Better `child_process` error handling for a couple scenarios with the finger pointing at it 👉
Also make sure we handle all of these scenarios:
1. Child process fork script throws an `uncaughtException` or `unhandledRejection`
- These are captured and serialized back to the parent and stored in `childErrors` and exposed if we never get a successful rendered HTML response.
2. Child process fails to startup
- Render process is rejected in the `child.on('error', ...` callback
3. 👉 Child process times out and is aborted
- Render process is rejected in the `child.on('error', ...` callback and any `childErrors` encountered are logged
4. 👉 Child process fork script throws an error in scope of in `process.on('message', async (renderOptions) => {`
- Child exits with code 1 and we reject the render process with the error
5. Child process exits with code 1 (error)
- Render process is rejected with any `childError` info
6. Child process exits with code 0 (success) but never sends back any HTML
- We have a `returnedData` data check and any child errors encountered are logged
We added the description in https://github.com/matrix-org/matrix-public-archive/pull/54 but then made a fix to one of the after render errors, https://github.com/matrix-org/matrix-public-archive/pull/57, which exposed that the description is then overriden when the powerlevel changes.
We now just override the getters directly instead of the internal fields which are overwritten internally.
---
This also fixes a warning from Hydrogen in the console:
```
disposable not found, did it leak? {kind: 'disabled', description: Array(3)}
```
Fix missing `room.observePowerLevels()` throwing error after render.
```
Uncaught (in promise) TypeError: this._room.observePowerLevels is not a function
at RoomViewModel2._recreateComposerOnPowerLevelChange (matrix-public-archive.js:23498:53)
at new RoomViewModel2 (matrix-public-archive.js:23469:14)
at mountHydrogen (matrix-public-archive.js:27131:25)
at matrix-public-archive.js:27174:1
```
Doesn't affect main-line rendering but the error is annoying to see. More Hydrogen boilerplate 😞
Use `DisabledComposerView` instead of no composer to add message:
> You're viewing an archive of events from 2022-02-08. Use a [Matrix client](https://matrix.to/#/!HBehERstyQBxyJDLfR:my.synapse.server) to start chatting in this room.
Also refactors some of the `views` and `viewmodels` out to their own files
Add test to make sure the archive doesn't fail when event for event relation is missing and not included in list of provided events. Like if someone is replying to an event that was from long ago out of our range.
In the case of missing relations, Hydrogen does `_loadContextEntryNotInTimeline` because it can't find the event locally which throws an `uncaughtException`. Before https://github.com/matrix-org/matrix-public-archive/pull/51, the `uncaughtException` killed the Hydrogen `child_process` before it could pass back the HTML. Now this PR mainly just adds a test to make sure it works.
```
TypeError: Cannot read properties of undefined (reading 'storeNames')
at TimelineReader.readById (hydrogen-web\target\lib-build\hydrogen.cjs.js:12483:33)
at Timeline._getEventFromStorage (hydrogen-web\target\lib-build\hydrogen.cjs.js:12762:46)
at Timeline._loadContextEntryNotInTimeline (hydrogen-web\target\lib-build\hydrogen.cjs.js:12747:35)
at Timeline._loadContextEntriesWhereNeeded (hydrogen-web\target\lib-build\hydrogen.cjs.js:12741:14)
at Timeline.addEntries (hydrogen-web\target\lib-build\hydrogen.cjs.js:12699:10)
at mountHydrogen (4-hydrogen-vm-render-script.js:204:12)
at 4-hydrogen-vm-render-script.js:353:1
at Script.runInContext (node:vm:139:12)
at _renderHydrogenToStringUnsafe (matrix-public-archive\server\hydrogen-render\3-render-hydrogen-to-string-unsafe.js:102:41)
at async process.<anonymous> (matrix-public-archive\server\hydrogen-render\2-render-hydrogen-to-string-fork-script.js:18:27)
```
Split off from https://github.com/matrix-org/matrix-public-archive/pull/43
Listen to `process.on('uncaughtException', ...)` and handle the async errors ourselves so it no longer fails the child process.
And if the process does exit with status code 1 (error), we have those underlying errors serialized and shown.
OpenTelemetry tracing so we can see spans where the app is taking time.
For the user, we specifically show the spans for the external API HTTP requests
that are slow (so we know when the Matrix API is being slow).
Enable tracing:
- `npm run start -- --tracing`
- `npm run start-dev -- --tracing`
What does this PR change:
- Adds OpenTelemetry tracing with some of the automatic instrumentation (includes HTTP and express)
- We ignore traces for serving static assets (just noise)
- Adds `X-Trace-Id` to the response headers
- Adds `window.tracingSpansForRequest` which includes the external HTTP API requests made during the request
- Adds a fancy 504 timeout page that includes trace details and lists the slow HTTP requests
- Adds `jaegerTracesEndpoint` configuration to export tracing spans to Jaeger
- Related to, https://github.com/matrix-org/matrix-public-archive/issues/26
We now run the Hydrogen render in a `child_process` so we can exit the whole render process. We still use the `vm` to setup the browser-like globals. With a `vm`, everything continues to run even after it returns and there isn't a way to clean up, stop, kill, terminate the vm script or context so we need this extra `child_process` now to clean up. I don't like the complexity necessary for this though. I wish the `vm` API allowed for this use case. The only way to stop a `vm` is the `timeout` and we want to stop as soon as we return.
Fix https://github.com/matrix-org/matrix-public-archive/issues/34
`renderHydrogenToString` is a pure function (probably) which means it will give the same output given the same input. This means, that if we give it a certain input and an error occurs, we should be able to reproduce it again if we have the arguments. This PR exposes those arguments in the logged error so we can investigate what's going wrong.
Added so we can investigate https://github.com/matrix-org/matrix-public-archive/issues/34 better and reproduce locally.
Fix image not running in K8s (ROSA)
Trying to solve the container exiting with error code `243`. `npm` throws `243` for any command you try to run with it.
Here is how to get into the container while the entrypoint normally errors.
```
$ oc run -i --tty --image=ghcr.io/matrix-org/matrix-public-archive/matrix-public-archive:sha-65edaea1a9713bb2cc93fa98638327fdeff765cd mpa-test2 --port=3050 --restart=Never --env="DOMAIN=cluster" --command /bin/bash
$ npm start
$ echo $?
243
```
Why does the same image work in Docker?
```
$ docker run -it --entrypoint /bin/bash ghcr.io/matrix-org/matrix-public-archive/matrix-public-archive:sha-65edaea1a9713bb2cc93fa98638327fdeff765cd
$ npm start
# it starts ✅
```
## Root cause
Seems like it could be either of these issues. Both are `8.6.0`+ problems as well which lines up with the `npm` version available in each base image.
- https://github.com/npm/cli/issues/4769
- https://github.com/npm/cli/issues/4996
- Specific comment about breaking stuff in k8s for other people, https://github.com/npm/cli/issues/4996#issuecomment-1150290804
Easy get the app running steps:
1. Install
2. Edit the config to point at your homeserver
3. Run the app!
---
This is made possible thanks to the change we made in https://github.com/matrix-org/matrix-public-archive/pull/20#discussion_r897567619 to start using a scoped custom version of `hydrogen-view-sdk`. This includes all of the scratch changes necessary to get this project running which makes the `npm install` just work out of the box without all of the `npm link` hassle.
1. Build test homeserver Docker images which can federate with each other
2. Run end-to-end (e2e) tests
#### Dev notes
Sharing variables across jobs when the `services` field can't access the `env` context, https://github.community/t/how-to-use-env-with-container-image/17252/24
```yaml
env:
FOO: bar
jobs:
set_env:
outputs:
var: ${{ steps.save_var.outputs.var }}
steps:
- id: save_var
run: echo "::set-output name=var::${{ env.FOO }}"
actual_job:
needs: set_env
container:
image: ...whatever_you_need_here...${{ needs.set_env.outputs.var }}
```
Remove `matrix-bot-sdk` usage in tests because it didn't have timestamp massaging `?ts` and it's not really necessary to rely on since we can just call the API directly 🤷. `matrix-bot-sdk` is also very annoying having to build rust crypto packages.
We're now using direct `fetch` requests against the Matrix API and lightweight `client` object.
All 3 current tests pass ✅
Eric Eastwood
