You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Should FunctionCallContent derive from UserInputRequestContent? #6831
If the intent of UserInputRequestContent is that it signals any content that the caller must respond to in a subsequent request, should FunctionCallContent derive from it? It currently doesn't.
I'm not sure that we should consider UserInputRequestContent as a call for all kinds of input. It is specifically "User" input, where user here means end user, not developer/caller. We could change it to be InputRequestContent though, which would be broader and in that case having FunctionCallContent derive from it as well would probably make sense.
We could change it to be InputRequestContent though
It does seem like we need some kind of indicator on content that it requires follow-up (and without said follow-up, subsequent requests will fail). I also think "user" is ambiguous and unnecessary.
Activity
westey-m commentedon Sep 22, 2025
I'm not sure that we should consider
UserInputRequestContentas a call for all kinds of input. It is specifically "User" input, where user here means end user, not developer/caller. We could change it to beInputRequestContentthough, which would be broader and in that case having FunctionCallContent derive from it as well would probably make sense.stephentoub commentedon Sep 22, 2025
It does seem like we need some kind of indicator on content that it requires follow-up (and without said follow-up, subsequent requests will fail). I also think "user" is ambiguous and unnecessary.
stephentoub commentedon Oct 9, 2025
I think we can leave it as-is.