ISSUES

# A list of bugs, desired features, possible changes to the system, etc.
# please list with newest issues at the top

# Note as of 5/1/15: we've moved to Issues list on Github; for the material, Chris wasn't sure whether we wanted to move these to the Github issues list

* (9/28/14) R's %% and C's fmod() differ when have, e.g., -0.8 %% 1 -- R gives 0.2, and C gives -0.8; one gross way to 'fix' would be to give an error if negative numbers are supplied

* (9/27/14 - possible "heavy debug/check" compilation mode.

* (7/24/14) Using 'equals(a,b)' in a model definition, when both a, b are constants, (correctly) attempts to evaluate 'equals(a,b)' at compile-time, but this errors-out because 'equals' refers to function in testthat package, rather than the intended 'nimbleEquals()' function.

* (7/10/14) Revisit use of `$` notation in the DSL.  This would break in normal R execution unless it is intercepted during nimbleFunction definition.

* (7/10/14) More profiling, particularly on C++ side

* (7/10/14) Possible bug from state-space model case: incorrect conjugacy determination.

* (7/10/14) 1D Laplace and AGQ

* (7/10/14) Make the inverse of as.matrix for a modelValues

* (6/27/14) using 'step' in data block of bugs model fails because we now use nimbleStep, so not backwards compatible... minor issue we may want to ignore so as not to mask R's step

* (6/22/14) add citations for our block RW (and maybe slice?) sampler algos in the manual

* (6/16/14) change scalar | and & to || and && ?

* (6/8/14) (CP) eventually release our C/R new distribution functions (d and r functions) as a separate R pkg?