Back to the main page.
Bug 3241 - improve the dashboard batch test scripts and resolve some regression errors
Status | ASSIGNED |
Reported | 2017-02-01 21:48:00 +0100 |
Modified | 2017-02-09 23:33:47 +0100 |
Product: | FieldTrip |
Component: | core |
Version: | unspecified |
Hardware: | PC |
Operating System: | Mac OS |
Importance: | P5 normal |
Assigned to: | Robert Oostenveld |
URL: | |
Tags: | |
Depends on: | |
Blocks: | |
See also: |
Robert Oostenveld - 2017-02-01 21:48:55 +0100
I put some effort into making a list of only passing test scripts, but nevertheless am encountering some test scripts that fail. I believe that might be due to some randomness on the environment with which jobs are started or on the compute nodes. These are the ones that will be reported in the automatic dashboard email tomorrow. 97dd149/test_bug1742.txt submitting job roboos_dccn_c021_p5492_b1_j001...=== test_bug1742 FAILED in 0 seconds 97dd149/test_bug2524.txt === test_bug2524 FAILED in 2 seconds 97dd149/test_tutorial_networkanalysis.txt === test_tutorial_networkanalysis FAILED in 220 seconds The first is a qsubcellfun test, which fails because matlab2016b cannot be found. I will change the run-test.sh script and add "module load cluster". The second calls fsl from the command line. I'll add "module load fsl". The third log file contains message: 'Unable to read file 'hdm'. No such file or directory.' it seems that it is not in the correct directory. The files are in /home/common/matlab/fieldtrip/data/ftp/tutorial/networkanalysis/. I will add that path to the test script with an explicit dccnpath.
Robert Oostenveld - 2017-02-01 22:03:32 +0100
(In reply to Robert Oostenveld from comment #0) this change is to the dashboard code: roboos@mentat002> git commit -a [master 79fdfd6] load the required modules, see http://bugzilla.fieldtriptoolbox.org/show_bug.cgi?id=3241 1 file changed, 4 insertions(+) this change is to the third FT test script mentioned above: roboos@mentat002> git commit -a [master ecf6001] ENH - go to the correct directory where the data is located prior to loading. See http://bugzilla.fieldtriptoolbox.org/show_bug.cgi?id=3241 1 file changed, 3 insertions(+) I hope this resolves it. The test run of this evening (linked to the ftp release) has already started. I'll start a separate schedule-tests.sh batch with the latest version (7bb1c24) of the master branch.