-
Notifications
You must be signed in to change notification settings - Fork 0
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
feat: automatically publish newly created flows #639
Conversation
version: 1 | ||
status: $status | ||
} | ||
const response: { insert_flows_one: { id: string; data: object } } = |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
My initial expectation was to destructure response
into something like insert_flows_one: insertFlowResponse
, but thinking about it more, do we keep response
as is so it's easier to track variables in publishFlow
lower down?
Crux of what I'm asking is, is it clearer to have:
await publishFlow(client, {
flow: {
id: response.insert_flows_one.id,
data: response.insert_flows_one.data,
},
publisherId: args.userId,
summary: "Created flow",
});
Or,
await publishFlow(client, {
flow: {
id: insertFlowResponse.id,
data: insertFlowResponse.data,
},
publisherId: args.userId,
summary: "Created flow",
});
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I simply extended the existing response
pattern used throughout this file because I was only adding a single variable, either syntax would ultimately accomplish the same thing 👍
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good to me, we're adding the Publish as part of creating a flow, with a summary note stating that. I'll look through the planx-new
change as well.
I only had one comment on the variable naming.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Something to take a look at here which might simplify a lot of the changes in theopensystemslab/planx-new#4230
One thing I'm not too sure of without a deep dive, is if there are places in e2e tests where the $admin
client is creating flows as opposed to a user with a JWT (therefore no x-hasura-user-id
session variable). If this is the case, we can insert using the $api
role instead which relied on a JWT 👍
@@ -148,48 +149,63 @@ export async function createFlow( | |||
name: string; | |||
data?: object; | |||
status?: FlowStatus; | |||
userId: number; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We can also infer this from the JWT making the request - this might both avoid a lot of the associated cascading changes from adding this new argument, as well as ruling out an incorrect value being passed in.
If we just use "x-hasura-user-id"
as a column preset, this will get read automatically from the JWT.
I'm pretty sure we can also toggle off access to the publisher_id
column once this is configured as well to make sure no incorrect values get written in.
Docs: https://hasura.io/docs/2.0/auth/authorization/permissions/column-presets/
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is a good shout - here's my understanding of how this is currently handled:
e2e/tests/api-driven
globalHelpers
all using$admin
client (create user, create team, create flow) which is why we're manually passing around thisuserId
demo-workspace
seems to be only exception to this which has uniquely defined/duplicated it's owncreateFlow
mutaton. Strangely/incorrectly though, this is also the only variation ofcreateFlow
I've found which doesn't create an associated operation and therefore I haven't added initial publishing support either (the tests never cover either of these scenarios)
e2e/tests/ui-driven
/helpers/context
similarly all using$admin
client (create user, create team, create flow, update settings/integrations, etc)- There is a
generateAuthenticationToken
method in this subdir to make a JWT, but the only reference I can find to this is when we call it once viaawait browserContext.addCookies()
- the future actions like creating a flow, updating flow status etc are still ultimately routed through an$admin
client afaik
Column preset definitely feels like the right direction of travel here, but also feels like an expansive piece of tech debt bigger than publishing alone ??
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yeah similarly just been looking into this!
The column preset is a great solution for all real-life applications, but doesn't marry up to how we're creating flows in e2e environments (set up using $admin
with no permissions, no user
record associated via JWT).
I don't think making the userId
param optional (for actual users) and conditionally reading argument or parsing from the JWT would make things any simpler - quite the opposite.
Agreed - this is a nice direction but a bigger change than I anticipated - this is a good route for now for sure ✅
This brings E2E helpers in sync in functionality introduced here: theopensystemslab/planx-new#4230