Skip to content

Conversation

@mzgubic
Copy link
Member

@mzgubic mzgubic commented Apr 29, 2021

Do we want this?

Closes JuliaDiff/FiniteDifferences.jl#149

@oxinabox
Copy link
Member

oxinabox commented Apr 29, 2021

I don't think we want this.
Special cases make it hander to maintain and use.
We have a type DoesNotExist which indicated to ignore it.
we should thus ensure when appropriate that type shows up.
we thus might want to normalize composites where all fields do not exist during the call to rand_tangent

@mzgubic
Copy link
Member Author

mzgubic commented Apr 29, 2021

Yeah fair. On the other hand I think @willtebbutt was reluctant to do a lot of special casing in rand_tangent

@oxinabox
Copy link
Member

I think special casing in rand_tangent is better.
Is a set of simpler methods.

If we did automatic simplification in it for structs it generated that were all DoesNotExist fields to become DoesNotExist then that would cover same as this but be nicer

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Should rand_tangent(2:3) return DoesNotExist()?

2 participants