[aspect-devel] aspect ctest fails?

Timo Heister heister at clemson.edu
Thu Jun 8 01:10:35 PDT 2017


I guess it is not documented in a great way, but the message you get
when you run with ASPECT_RUN_ALL_TESTS=OFF already explains this, see
https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_geodynamics_aspect_blob_master_tests_cmake_print-5Ftest-5Finfo.sh&d=DwIBaQ&c=Ngd-ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=c08Btfq4m9QEScXN3ZQwLZzzWQE7S8CYq1IYuzKV_Zk&m=CNH1WbO-rmy22jRgwLLXpOTbcae18m8rcvCJXmYNHOw&s=200hys7sVkvuy1aC2INAmuforzCfQlJQdo9vVfbJ4Co&e= 

> @Timo: Is it possible to add more
> tests to the simple tests and maybe also decrease their sensitivity
> (i.e. increase the tolerance for numdiff)?

Certainly and I think this would be great to have! You just need to
put the keyword "QUICK_TEST" inside the prm and it will be picked up
immediately. It is a bit harder to make the tests very robust. Just
changing tolerances for numdiff won't work (take x or x+1 Stokes
iterations for example).


-- 
Timo Heister
http://www.math.clemson.edu/~heister/


More information about the Aspect-devel mailing list