As a consequence, increase the tolerance for a test with mfs=3 (it now uses the
same tolerance as the corresponding test for mfs=2).

Closes: #1883

Closes: website#3

– Drop dynamic_g1_nz.m generated file
– Fix size of dynamic derivatives (legacy representation) when there are unused
  exogenous and “nostrict” option is given
– Block decomposition: another fix for sparse indices of static Jacobian when
  mfs > 0

This can prevent mistakes when the array is modified.

Ref. #1859

missing_observations_kalman_filter.m: provide debugging information on PKF problems

See merge request Dynare/dynare!2122

As a consequence, fix steady state computation when mfs > 0.

E.g. indentation was incorrect following “model(block);”.

[skip ci]

Bytecode w/ block decomposition: no longer output derivatives w.r.t. exogenous and endogenous outside the block

It should probably have been removed in 9c6e2199.

 - Use bsxfun for centering data if possible,
 - Factorise LU decomposition,
 - Remove useless operations during the presampling step.

Ref. #1859

We also take advantage of the fact that the non-block version is always
available next to the block one, so when we are only interested in the residual
as a whole, we simplify by using the non-block version.

This workaround already existed for the legacy representation.

dsge_likelihood.m: flag analytical_derivation as incompatible with endogenous_prior

See merge request Dynare/dynare!2120

As a consequence, a new “block_decomposed” option of the bytecode MEX has been
introduced to explicitly select the block-decomposed version.

Note that we do not always use the “block_decomposed” option even when the
“block” option has been passed to the user, in situations where the block
decomposition brings nothing (e.g. when evaluating the residuals of the whole
model).

[skip ci]

Ref. !2118