add mcmc test with some RHSonly nodes

[perrydv]

I don't think any of the MCMC tests use a model with RHSonly nodes

[danielturek]

Surely one of more of the standard BUGS examples (which are tested in test-mcmc) have RHS only nodes?

If not that, then the testing for the elliptical slice sampler, and some of the MCMC tests for the block sampler in test-mcmc appear to have RHS only precision / covariance matrices.

Is that what you'te looking for @perrydv ? If not, the I'm missing the gist of this.

[perrydv]

I haven’t looked further, but I think for the BUGS examples we typically provide constants for what would otherwise become RHSonly nodes, but once the constants are baked in, the RHSonly status disappears because they are not longer nodes at all. (That will change on newNodeFxns, where non-scalar constants are no longer baked in.)

I did something yesterday that caused a bug for RHSonly nodes (now fixed), yet it passed test-mcmc completely. That’s how I concluded that test-mcmc must lack any RHSonly cases.

On Jun 3, 2016, at 5:07 PM, Daniel Turek [email protected] wrote:

Surely one of more of the standard BUGS examples (which are tested in test-mcmc) have RHS only nodes?

If not that, then the testing for the elliptical slice sampler, and some of the MCMC tests for the block sampler in test-mcmc appear to have RHS only precision / covariance matrices.

Is that what you'te looking for @perrydv https://github.com/perrydv ? If not, the I'm missing the gist of this.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/nimble-dev/nimble/issues/171#issuecomment-223722276, or mute the thread https://github.com/notifications/unsubscribe/AHpCMcPRSvAu235aToRcI1ycZCmOafOwks5qIMG-gaJpZM4It5G8.