-
Notifications
You must be signed in to change notification settings - Fork 43
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
StackOverflowError during type inference on PDMats 0.11.5 #150
Comments
Maybe the same error as in invenia/KeyedDistributions.jl#24? |
Seems likely; this error also does not happen on Julia 1.7. |
Changing Line 4 in 15fecad
struct PDiagMat{T<:Real,V<:AbstractVector} <: AbstractMatrix{T} fixes the issue for me.
I assume that 1.6 (but apparently not Julia 1.7) has a problem with the type-inference failures of |
Ah yes, sorry I should have looked there beforehand. It seems there are different subtle changes that can trigger the type-inference failure, but PDMats is not at fault given it can also happen without PDMats. Feel free to close this issue. |
Also the other (well, actually same) issue in KeyedDistributions is caused by the change of the type parameter in PDMats. I'd like to understand what's going on there and why it causes problems with Julia 1.6 since it is the correct and desired parameterization. However, maybe it would be easier to just apply the fix mentioned above - even though I think it should not be needed... |
I'm not familiar with type inference so not sure what is going on, and whether this has more to do with
KeyedDistributions
than withPDMats
. However I wasn't seeing the error with PDMats 0.11.4, but now seeing it with 0.11.5.Here's my MWE below, happy to provide more information.
1.log
2.log
3.log
Here's the StackOverflowError during type inference (from 1.log, abriged with
...
for the 3 lines that get repeated for a while):The text was updated successfully, but these errors were encountered: