http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59595
--- Comment #3 from dave.anglin at bell dot net --- I have been trying to narrow down the change that caused the regression. r205921 is OK and r205955 is bad. Most recent build that I have noted as bad is r206196. Last test results from box are here: http://gcc.gnu.org/ml/gcc-testresults/2014-01/msg02299.html From the above, it looks like the failure points for 59913 and 59595 are different. Dave