Update Known bugs present in the current stable version authored by Sébastien Villemot's avatar Sébastien Villemot
This page documents known bugs in Dynare stable. This page documents known bugs in Dynare stable.
* Perfect foresight simulation with `bytecode` and `stack_solve_algo=1` may fail to converge (fixed in 67a003aa730e59f7b254353ca124909d69dcaaca)
* Perfect foresight simulation with `block` (but without `bytecode`) of blocks of type “Solve two boundaries simple/complete” give wrong results in the last simulation period (fixed in 0147faad5f03cc641fd22d3e5733d4862964ef0b) * Various problems with perfect foresight simulations in combination with `block` and/or `bytecode` options:
* Steady state computation with `bytecode` and `block` could fail if some equations were marked `[static]` (if temporary terms were created for the derivatives) (fixed in preprocessor@b78ac1d31fd27b8cc89ac2162d58d029821f9af1 and 8f5815d01d3604494f4a0bffc5f92d842972b844) + Simulation with `bytecode` and `stack_solve_algo=4` of *linear* “Solve two boundaries simple/complete” blocks can give incorrect results (fixed in 65e18fe8834b31a885667e0e498e3a2ff0f1f47a)
* Steady state computation with `bytecode` and `block`, and `solve_algo` ⩽ 4 or ⩾ 9 can needlessly fail + Simulation with `bytecode` and `stack_solve_algo=1` may fail to converge (fixed in 67a003aa730e59f7b254353ca124909d69dcaaca)
* Perfect foresight simulation with `bytecode` and `block` of *linear* “Solve forward simple/complete” blocks will give incorrect results (fixed in 863fa2de1c04d2a33aef042adacdb78dae4d0ae1) + Simulation with `block` (but without `bytecode`) of blocks of type “Solve two boundaries simple/complete” gives wrong results in the last simulation period (fixed in 0147faad5f03cc641fd22d3e5733d4862964ef0b)
* Perfect foresight simulations with block decomposition (and without `bytecode` option) will crash or give incorrect results if there are blocks of type “Solve forward/backward simple/complete” (fixed in fa7b331cf4143816a6565e3b5075cc3b10ba879d) + Simulation with `bytecode` and `block` of *linear* “Solve forward simple/complete” blocks will give incorrect results (fixed in 863fa2de1c04d2a33aef042adacdb78dae4d0ae1)
* Steady state computation with `bytecode` and `block` options and `solve_algo=6` will crash or give incorrect results if there are blocks of type “Solve forward/backward complete” (fixed in 188df7711d4f4a704d7768a61ce70d492cfec9a3) + Simulation with block decomposition (but without `bytecode` option) will crash or give incorrect results if there are blocks of type “Solve forward/backward simple/complete” (fixed in fa7b331cf4143816a6565e3b5075cc3b10ba879d)
* Perfect foresight simulations with `bytecode` and `block` options and `stack_solve_algo={0,1,4}` will crash or give incorrect results if there are blocks of type “Solve forward/backward complete” (fixed in cdb4f872b24dd811c6053026695d342e05d2b1f9 and 188df7711d4f4a704d7768a61ce70d492cfec9a3) + Simulation with `bytecode` and `block` options and `stack_solve_algo={0,1,4}` will crash or give incorrect results if there are blocks of type “Solve forward/backward complete” (fixed in cdb4f872b24dd811c6053026695d342e05d2b1f9 and 188df7711d4f4a704d7768a61ce70d492cfec9a3)
* Perfect foresight simulations with block decomposition (and without `bytecode` option) will give incorrect results if a block of type “Solve backward simple/complete” is in the model (purely forward equations that cannot be simply evaluated) (fixed in bb793d87cc7dc35fe4fd4767322c2f7b57788eb9) + Simulation with `block` (but without `bytecode`) gives incorrect results if a block of type “Solve backward simple/complete” is in the model (purely forward equations that cannot be simply evaluated) (fixed in bb793d87cc7dc35fe4fd4767322c2f7b57788eb9)
* Various problems with steady state computation in combination with `block` and `bytecode` options:
+ Steady state computation with `bytecode` and `block` could fail if some equations are marked `[static]` (if temporary terms were created for the derivatives) (fixed in preprocessor@b78ac1d31fd27b8cc89ac2162d58d029821f9af1 and 8f5815d01d3604494f4a0bffc5f92d842972b844)
+ Steady state computation with `bytecode` and `block`, and `solve_algo` ⩽ 4 or ⩾ 9 can needlessly fail
+ Steady state computation with `bytecode` and `block` options and `solve_algo=6` will crash or give incorrect results if there are blocks of type “Solve forward/backward complete” (fixed in 188df7711d4f4a704d7768a61ce70d492cfec9a3)
* The `check` command will crash or give incorrect results when using block decomposition and a block of type “Solve backward complete” is in the model (several interrelated purely forward equations) (fixed in dd34554c4e23a12159388d54fe304698c4d64ca5) * The `check` command will crash or give incorrect results when using block decomposition and a block of type “Solve backward complete” is in the model (several interrelated purely forward equations) (fixed in dd34554c4e23a12159388d54fe304698c4d64ca5)
* The options `static` and `incidence` of the `model_info` command do not work as advertised in the reference manual (fixed in 685598c6e2690cfbc4e677cd2148387716ef8243 and preprocessor@7df67819821ac5946e1b4ace3e974752ce533827) * The options `static` and `incidence` of the `model_info` command do not work as advertised in the reference manual (fixed in 685598c6e2690cfbc4e677cd2148387716ef8243 and preprocessor@7df67819821ac5946e1b4ace3e974752ce533827)
* Method of moments will crash if no `matched_moments`-block is present (bug fixed, see !1997) * Method of moments will crash if no `matched_moments`-block is present (bug fixed, see !1997)
... ...
......