Dynare "make check" failed tests

Installation and Setup
1

Building dynare 4.6.4 from source the following tests fail after running: “make check”

================================
DYNARE MAKE CHECK MATLAB RESULTS

| TOTAL: 1012
| PASS: 990
| FAIL: 12
| XFAIL: 10
| XPASS: 0
|
| LIST OF FAILED TESTS:
| * particle/first_spec.mod
| * particle/local_state_space_iteration_k_test.mod
| * …/matlab/missing/stats/gamrnd.m#1
| * …/matlab/missing/stats/gamrnd.m#2
| * …/matlab/missing/stats/gamrnd.m#3
| * …/matlab/missing/stats/gamrnd.m#4
| * …/matlab/missing/stats/gamrnd.m#5
| * …/matlab/missing/stats/gamrnd.m#6
| * …/matlab/missing/stats/gamrnd.m#7
| * …/matlab/missing/stats/gamrnd.m#8
| * …/matlab/modules/dseries/src/@dseries/std.m#3
| * …/matlab/modules/dseries/src/@dseries/std.m#4
|
| LIST OF 10 SLOWEST TESTS:
| * k_order_perturbation/burnside_k_order.mod - 1170.139263
| * observation_trends_and_prefiltering/MCMC/Trend_no_prefilter_MC.mod - 510.943656
| * observation_trends_and_prefiltering/MCMC/Trend_loglinear_prefilter_MC.mod - 489.904405
| * observation_trends_and_prefiltering/MCMC/Trend_prefilter_MC.mod - 469.613683
| * estimation/tune_mh_jscale/fs2000.mod - 435.222545
| * observation_trends_and_prefiltering/MCMC/Trend_loglinear_no_prefilter_MC.mod - 427.674094
| * shock_decomposition/ls2003_plot.mod - 299.502505
| * gsa/ls2003a.mod - 267.550380
| * gsa/ls2003.mod - 253.940911
| * TeX/fs2000_corr_ME.mod - 163.182025

Makefile:1411: recipe for target ‘check-matlab’ failed
make[2]: *** [check-matlab] Error 1
make[2]: Leaving directory ‘/path/dynare/build/dynare-4.6.4/tests’
Makefile:895: recipe for target ‘check-am’ failed
make[1]: *** [check-am] Error 2
make[1]: Leaving directory ‘/path/dynare/build/dynare-4.6.4/tests’
Makefile:422: recipe for target ‘check-recursive’ failed
make: *** [check-recursive] Error 1

Any reasons why these 12 tests would fail?

2

The only way to find out is to have a look at the artifacts generated. Only the first two are kind of relevant. The others suggest that there is potentially something with the statistics toolbox or functions going on.

3

The first 2 tests fail because the source tar file seems to be missing the tests/particle/first_spec_common.inc file. Was it excluded from the tar for a specific reason?

4

That is strange. How did you download the program and which OS are you using?

5

I downloaded the source tar file from Download | Dynare under the source tab.

I reran the tests and copied the tests/particle/first_spec_common.inc from the dynare repository tests/particle/first_spec_common.inc · master · Dynare / dynare · GitLab. The first two tests did not fail after that. I also copied matlab/missing/stats/gamrnd.m · master · Dynare / dynare · GitLab file over and the gamma distribution tests also did not fail.

However I still have 4 failing tests:

| * …/matlab/modules/dseries/src/@dseries/mean.m#3
| * …/matlab/modules/dseries/src/@dseries/mean.m#4
| * …/matlab/modules/dseries/src/@dseries/std.m#3
| * …/matlab/modules/dseries/src/@dseries/std.m#4

These tests seem to be commented out n the files that come packaged with the source tar. Do we know if that is why they are failing?

6

I opened a ticket at Investigate reasons for incomplete source package distributed via Webpage (#1803) · Issues · Dynare / dynare · GitLab
I think you can safely ignore the last four tests.

7

Thank you very much! Much appreciated!