Perry de Valpine

`model[[nodes]] <- scalar` breaks

1

This used to compile (whether it gave correct behavior would need to be checked). Reproducible example: ``` nf

If an element in the list provided to `model$setInits` is the wrong size, it will not be used and only an obscure warning message from C++ will be seen by...

I noticed that `SEXP_2_int` in the C++ code uses the `offset` argument inconsistently for double vs integer or logical input. I want to check if this is correct.

allow non-abstract do-nothing default methods in nimbleFunctionVirtual

10

This is not for the imminent release. @danielturek This PR provides the features needed to modify #1181 so that `sampler_BASE` could be expanded rather than adding a `sampler_BASE2`. Let's see...

This test file dates back to a change in the compilation system. It re-runs a bunch of math tests to compare old and new systems. However the actual math tests...

When one nimbleList contains another, and the contained nimbleList is a predefined, the dll in which we look for SEXP-functions for accessing the C++ objects is the wrong dll. It...

nodeFunction problem when a for loop index is also a variable in the model

1

A user found this impossible to pin down and reported that it works in jags. Here is a minimal reprex: ``` library(nimble) mc

This was raised by a user and makes a lot of sense.

Support for `ifelse`

1

For simplicity should we add support for `ifelse`? This contributed to user error in transitioning from JAGS to NIMBLE in a nimble-users list message dated 31 July 2021.

Suggestion to error-trap variable names rather than node names in values()

2

The code ``` values(model, nodes)