-
Notifications
You must be signed in to change notification settings - Fork 132
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
Question: customConfig vs hostManifests in v2 #900
Comments
I have tried to use this, but since the apps don't have access to this via their metadata, it seems like it's redundant. It would be really handy if this was a way for the app directory to pass config to the app. I would be very pleased to see this added to Some use cases:
|
@robmoffat as discussed at the SWG meeting Also as discussed at the meeting, I'm aware that our product and others have their own mechanisms for creating custom config, specified in the host manifest, and for an app to retrieve it (as outlined in the issue). However I'm also keen for the standardized part of the appD record to handle as much as it can without a hostManifest... @pgn-vole any further thoughts from you given this discussion? |
@kriswest yes seems sensible that other apps can't access it. |
It would be interesting to understand live uses cases of customConfig that cannot be handled via hostsManifests. In our case all customConfig keys (that were application launcher specific), have been moved to launcher specific hostManifests. So only the hostManifest is used to convey metadata from application to launcher. Ability for a launcher to pass metadata to applications is interesting use case, I would however see such environment specific properties exposed via dedicated fdc3 method rather than having an application directory extending applications metadata at runtime. |
I suspect that there are use cases for app implementations that render different content based on config - e..g. visualizations of industry sectors, pricing tiles etc.. Hence, one use case may be the passing of configuration to an app independent of a Further, the hostManifest is information passed to the Desktop Agent, which only passes part of it on to application(s): the app itself through Whereas, we could choose to expose customConfig to the application itself through the FDC3 API, perhaps via another field in the (just thinking out loud about potential use cases, rather than specifically advocating for them) |
At the last two SWG meetings (#913 #934) I provided a summary of this issue and asked if anyone wanted to speak for or against the proposed change to make customConfig accessible to apps - no one has taken me up on that offer. Hence, I suggest we proceed with the original suggestion to deprecate |
At SWG meeting #1005 a number of participants spoke up for a use case for custom configuration that could be retrieved by an application using a standardized function, where the config might affect the behaviour of the app (and do so in vendor-agnostic fashion). E.g. a market sector visualization or blotter might be filtered by region (APAC, EMEA, etc.). This is a different use case (intended for apps, vendor agnostic format) to that of the Hence, it was agreed to close this issue and the associated PR in favour of a new issue to properly serve this use case. Closing in favour of #1006 |
At meeting #1013 it was agreed that
|
Question Area
[x] App Directory
[ ] API
[ ] Context Data
[ ] Intents
[ ] Use Cases
[ ] Other
Question
With the introduction of
hostManifests
property what is the purpose of thecustomConfig
property from now on?The API doc of
customConfig
states:However, this seem to have the same purpose as the hostManifests property because the hostManifests can now too contains launcher specific configuration/metadata entries.
From my point of view I always found the customConfig property a bit odd because it was felt like bag of launcher specific metadata that could easily conflict between each others if 2 launchers for instance rely on a customConfig property of the same name. The
hostManifest
feels like a much more elegant solution at replacing that kind of use case.The text was updated successfully, but these errors were encountered: