TODO:
– add test that checks that computed exos actually give the requested endos
  when doing a regular simulation
– add test to check that a given exo is not simultaneously in
  perfect_foresight_controlled_paths and shocks, or that a given exo/endo does
  not appear two times in perfect_foresight_controlled_paths for a given period
– support perfect_foresight_with_expectation_errors (+test)
– support homotopy (+test) (+linearization?)
– support LBJ (+test)
– support stack_solve_algo=7 (reusing function for stacked Jacobian substitution)
– integrate all tests in meson
– documentation

[skip ci]

 discretionary_policy: fix noprint option

See merge request !2349

Output argument and return statement were missing

 fix loop in kalman_update_engine.m if regime_start hits 0

Closes #1956

See merge request !2348

Closes #1956

 model_diagnostics.m: set missing variable for bytecode

See merge request Dynare/dynare!2347

[skip ci]

[skip ci]

In Octave, calling @dseries/plot multiple times with `hold on` does
not allow for plotting multiple time series of different lengths on
the same figure (it throws an error).

[skip ci]

[skip ci]

 conditional forecasts: allow using vectors in blocks

Closes #1949

See merge request Dynare/dynare!2346

Closes #1949

Ref. !2344

Ref. !2343

Ref. !2342

Closes #1953

This partially reverts commit af274aec

Add integration test for newrate with analytic_derivation

Closes #1935

See merge request Dynare/dynare!2340

Related to #1935

fix to #1935.

See merge request Dynare/dynare!2339

fix to #1935. Forced options_.analytic_derivation=-1; must be set before entering dynare_minimize_objective

It should have been done in commit b9af92eb.

By the way, fix a font decoration issue.

[skip ci]

 fix identification bugs

See merge request Dynare/dynare!2333

See commit 2983aafa for the rationale.

In situations where the solution cannot be found for an informational period > 1
using the same completion share as for the first informational period, suggest
to decrease the value of the homotopy_max_completion_share option.

[skip ci]

Use the model simulation based on a first-order local approximation as
the initial guess for the nonlinear solver in each period. Can
speed-up extended path simulations.

We could also consider simulations based on higher order
approximations, but I did not find any significant improvements with a
second order approximation.