Release Notes 2.36

cmdstan.txt

 - No longer force fix_param when model has no parameters. (#1288)

 - Updated stansummary to use split, rank-normalized Rhat and ESS. (#1301)

 - Fixed stansummary crashing if it is run on a csv file without any `__` columns (#1283)

 - Fixed stansummary not rebuilding when a dependent file changes (#1283)

 - Removed support for adding makefile variables in `~/.config/stan/make.local`. Use the `make/local` file in the CmdStan directory instead. (#1282, #1284)

 - Windows: Don't try to copy tbb.dll if TBB_TARGETS is undefined. (#1285)

math.txt

Breaking changes:

 - Removed deprecated `_log()` probability functions. (#2995)

 - Moved constraint transforms to their own folder. (#3039)

 - Updated arena_matrix to use move semantics. (#2928) 
 
 - Un-vendored a great many unused boost libraries. (#3129)
 
 - C++17 is now required.
 
 Other changes:
 
 - Added `sum_to_zero_constrain`, `sum_to_zero_free`, and `check_sum_to_zero`. (#3101)

 - Added beta_neg_binomial_lpmf. Allows modeling heavy-tail count data. (#3108)

 - Added beta_neg_binomial_lccdf. (#3114, #3118)

 - Added beta_neg_binomial_lcdf. (#3116)

 - Added beta_neg_binomial_cdf. (#3120)

 - Added beta_neg_binomial_rng. (#3126)

 - Added overloads for tuples of bounds to `lub_free` and `lub_constrain`. (#3087)

 - Added error checking functions for row/column stochastic matrices. (#3095)
 
 - Added constructor to vari for passing initial values and adjoints. (#3089)
 
 - Improved efficiency of `finite_diff_gradient_auto`. (#3096)
 
 - Functions concerned with the number of entries in a container (`size`, `cols`, `rows`, ...) now consistently use `long int` as their return type. (#3086)
 
 - Removed support for adding makefile variables in `~/.config/stan/make.local`. Use the `make/local` file in the Math directory instead. (#3090)

 - Fixed an indexing bug inside gp_matern32_cov. (#3084)

 - Fixed the new `wiener_lpdf` overloads evaluating their expression template arguments more than once. (#3112)

 - Fixed some failures to compile calls to `pow` when using libc++ 19. (#3110)

 - Updated templated complex functions in the Stan math library to be better candidates than the standard library functions during ADL. (#3122)

 - Fixed a warning about a mismatched struct/class declaration from clang 18. (#3091)

 - Fixed compilation of `fabs` with Apple Clang 15. (#3093)

 - Fix Eigen issue [2852](https://gitlab.com/libeigen/eigen/-/issues/2852) so that Eigen's kronecker product type works  with `value_type_t`. (#3124) 

stan.txt

 - Added de/serialization for the `stochastic_row_matrix` and `stochastic_column_matrix` types. (#3298, #3304)

 - Added de/serialization for the `sum_to_zero_vector` type. (#3308)
 
 - Relaxed restrictions on HMC to allow running when a model has no parameters. (#3071, #3303)

 - Updated `rvalue()` functions that use `multi_index` to return an Eigen expression of multi indexing instead of eagerly creating a new matrix from the indices. (#3250)

 - Improved error messages when user-specified initializations fail. (#3291)

 - Fixed progress output formatting being inconsistent for the non-adaptive samplers. (#3296)

 - Improved parsing of Stan CSV files to allow for getting MCMC sample from fixed_param and samples with saved warmup draws and to be able to correctly get ADVI sample. (#3311)
 
 - Improved parsing of Stan CSV files when they contain nan or infinite values. (#3318)
 
 - Added a new `stan::mcmc::chainset` object, exposing split-Rhat, split-ESS and Monte Carlo standard error diagnostics. This replaces the `chains<>` object. (#3312, #3313)

 - Fixed a header include order issue that caused compilation with `-DSTAN_MODEL_FVAR_VAR` to fail for certain models. (#3294)

 - Removed support for adding makefile variables in `~/.config/stan/make.local`. Use the `make/local` file in the Stan directory instead. (#3295)

stanc3.txt

 - Removed code following the last round of syntax deprecations that is currently scheduled. The final one (use of `real` values in conditionals) is now a type error and can no longer be automatically canonicalized. (#1420)

 - Added the `jacobian +=` statement, an analogue of `target +=` but which can be conditionally disabled by Stan's algorithms when a unadjusted estimate is required. (#1435, #1437, #1471)

 - Added `stochastic_row_matrix` and `stochastic_column_matrix` transforms for matrix types. (#1442)

 - Added the `sum_to_zero_vector` transform for vector types. (#1443)

 - Added distribution: beta_neg_binomial, with corresponding lpdf, cdf, lcdf, lccdf, and rng functions. (#1452, #1459, #1467)

 - Added all-vector signatures to the newer versions of `wiener_lpdf`. (#1451)

 - `stanc.js` can now accept models which contain `#include` statements. A fourth argument is available on the javascript stanc function which must be either `undefined` or a object mapping included file names to Stan source code as strings. (#1433)

 - Enabled OpenCL support for the `binomial_logit` and `binomial_logit_glm` distributions. (#1368)

 - Fixed an issue where certain overloads of distributions could lead to the `~` statement producing uncompilable C++. (#1466, #1474)
 
 - Fixed an issue where `operator/` was not generating the correct C++ for complex linear algebra types. (#1421)
 
 - Fixed an issue where certain variable names would end up conflicting with system-specific macros in the generated C++. (#1429)
 
 - Fixed an issue where void functions were not given the same checks as returning functions, leading to uncompilable C++ in some cases. (#1471)
 
 - Updated parameter code generation to use `auto` so that the parameter types for matrices can be Eigen maps. (#1441)
 
 - Rolled back js_of_ocaml to 5.4.0 for compatibility with QuickJS. (#1427)

 - Added builds of `stanc.js` with pretty-printing and debug info. (#1440)