> Apparently in masking sage-4.4 to revert to 4.3.5 I had neglected to > mask sage-doc-4.4. I had the wrong sage-doc installed for the test! In > reverting to 4.3.5 other test failures were due to scipy as you > indicate and others were due to matplotlib. I now have been able to > reproduce test results I had previously for 4.3.5. Upon then upgrading > to 4.4 I have many failures of the "type" indicated in the first post > of this thread; i.e. of the "type" > > *** glibc detected *** /usr/bin/python2.6: corrupted double-linked list: > > with an associated backtrace and memory dump. I suppose Christopher > will observe the same on his amd64. For the record on x86 I have a number of tests failing due to maxima-5.21.1/ecls-10.4. Hopefully those will go away in sage-4.4.1. Most of the other failures are old. I do not have anything like you do. What is worrying is I am guessing you don't see any patterns in the failures. If it is all over the place I would have a serious look at sage-lib and sage-core.
Francois
