As far as I understand the test the problem is the following:
In this case the test defined SuiteSparse_UMFPACK_FOUND which activates the test.
On the other hand HAVE_SUITESPARSE_UMFPACK which is checked in our umfpack.hh is set to ENABLE_SUITESPARSE in config.h. This, however, is only set if SuiteSparse_FOUND is true.
Designs
Child items 0
Show closed items
No child items are currently assigned. Use child items to break down this issue into smaller parts.
Linked items 0
Link issues together to show that they're related.
Learn more.
The AMG problem is AMG-inherent: It is AMGs policy to do so if neither UMFPack nor SuperLU is found on the system.
I am all for discussing whether requiring one of these for AMG is a good idea.
But to avoid confusion: This is not just a problem of not skipping the test. (I have already patched it locally to do so.) The main problem is that the results of the suitesparse test are inconsistent.
Yes, as expected, the build does no longer fail when properly skipping the test. (My local patch did the same.) However, this bug was only the symptom of the underlying problem that the results of the suitesparse test are inconsistent. This issue is still present.
Another symptom of the same problem is, that the umfpackamgtest still compiles and runs but without using umfpack, although umfpack is installed.