Re: Build failed in Jenkins: Stan Distribution Tests #475

[Marcus Brubaker]

This looks like a stale file from the previous build.  It may need some manual poking.

On Fri, Jan 3, 2014 at 8:21 PM, <mc.sta...@gmail.com> wrote:

See <http://d1m1s1b1.stat.columbia.edu:8080/job/Stan%20Distribution%20Tests/475/changes>

Changes:

[Daniel Lee] Test mdivide_{left|right}_ldlt and log_determinant_ldlt.

[Daniel Lee] Split trace_gen_quad_form tests into their own files.

[Daniel Lee] Add unit tests for trace_inv_quad_form_ldlt and fix derivative error.

[Daniel Lee] Rename trace_inv_quad_form_ldlt with three arguments to trace_gen_inv_quad_form_ldlt to be consistent with the non-inverse form of the trace-of-quadratic-form functions.

[Daniel Lee] Rename test functions to be consistent.

[Daniel Lee] Unit test for trace_gen_inv_quad_form_ldlt.

[Daniel Lee] Expand testing for multi_student_t_log.

[Daniel Lee] Use LDLT_factor in multi_normal_log again now that the gradients are fixed.

[Daniel Lee] Use trace_inv_quad_form_ldlt() in multi_student_t.

[Daniel Lee] Fix function call in multi_gp_log.

[Daniel Lee] Fix memory issue with log_determiant_ldlt_test.cpp

[Daniel Lee] split apart src/stan/math/matrix/ldlt.hpp into 6 header files; created test files for the missing ones (no tests included)

[Daniel Lee] split src/stan/agrad/rev/matrix/ldlt.hpp into 6 header files; created test files for missing tests (no tests included)

[Daniel Lee] added failing finite diff test for agrad mdivide_left_ldlt

[marcus.a.brubaker] Fix LDLT finite difference tests.

[Daniel Lee] added more tests for log_determinant

[Daniel Lee] added test for trace_gen_inv_quad_form_ldlt using basic math and auto-diff

[Daniel Lee] changed include order of multiply

[Daniel Lee] testing mdivide_right_ldlt. tests fail

[Daniel Lee] adding math test for trace_inv_quad_form_ldlt

[Daniel Lee] added tests for trace_inv_quad_form_ldlt for gradients checking against auto-diff implementation

[marcus.a.brubaker] Fix mdivide_right_ldlt unit test.

[marcus.a.brubaker] use mdivide_right_spd instead of inverse to test gradients.

[bg2382] add some code comments

[marcus.a.brubaker] Add some comments on how LDLT_factor should be used.

[bg2382] add .vectorD() method to LDLT

[bg2382] use check_ldlt_factor() in distributions

[bg2382] forgot to commit check_ldlt_factor.hpp

[marcus.a.brubaker] Change types to silence new compiler warnings.

[marcus.a.brubaker] ASSERT_TRUE on LDLT_factor::success() in unit tests.

[marcus.a.brubaker] Forgot to commit this file, silences warnings.

[mbrubake] Add some more API documentation to LDLT_factor and LDLT_alloc.

[marcus.a.brubaker] Unit test for LDLT_alloc.

[Daniel Lee] adding more tests for LDLT_alloc

[Daniel Lee] Adding LDLT_factor test that breaks. Not just fails, but breaks

[marcus.a.brubaker] Check that a factorization attempt has happend in success()

[marcus.a.brubaker] Expand documentation to make it clear that you must call compute() if you use the default constructor.

[Daniel Lee] added stronger language indicating that success() must be true before using the other functions within LDLT_factor

------------------------------------------
Started by upstream project "Stan Header Tests" build number 140
originally caused by:
 Started by upstream project "Stan Unit Tests" build number 1139
 originally caused by:
  Started by GitHub push by syclik
[EnvInject] - Loading node environment variables.
Building on master in workspace <http://d1m1s1b1.stat.columbia.edu:8080/job/Stan%20Distribution%20Tests/ws/>
Fetching changes from the remote Git repository
Fetching upstream changes from https://github.com/stan-dev/stan.git
Checking out Revision b63bfa76cae32b599e5e7d925588ebbd54c6a04c (origin/develop)
Run condition [Current build status] enabling prebuild for step [Execute Windows batch command]
[Current build status] check if current [SUCCESS] is worse or equals then [FAILURE] and better or equals then [FAILURE]
Run condition [Current build status] preventing perform for step [Execute Windows batch command]
[workspace] $ cmd /c call C:\Users\daniel\AppData\Local\Temp\hudson4847282338798069688.bat

<http://d1m1s1b1.stat.columbia.edu:8080/job/Stan%20Distribution%20Tests/ws/>make> test-distributions -j4
uname: not found
uname: not found
INFO: Could not find files for the given pattern(s).
make: *** No rule to make target `src/stan/agrad/rev/matrix/ldlt.hpp', needed by `src/stan/model/model_header.hpp.d'.  Stop.
make: *** Waiting for unfinished jobs....

<http://d1m1s1b1.stat.columbia.edu:8080/job/Stan%20Distribution%20Tests/ws/>exit> 2
Build step 'Execute Windows batch command' marked build as failure
[WARNINGS] Skipping publisher since build result is FAILURE
Recording test results

--
You received this message because you are subscribed to the Google Groups "stan-buildbot" group.
To unsubscribe from this group and stop receiving emails from it, send an email to stan-buildbo...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Daniel Lee]

I just cleaned the workspace and triggered the build again.

I tried to be clever and didn't clean all before every build, but that doesn't work when files go missing.

I'll update it when I have internet from my computer.

You received this message because you are subscribed to the Google Groups "stan development mailing list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to stan-dev+u...@googlegroups.com.