
22 Oct
2013
22 Oct
'13
12:45 p.m.
Hi, in remove_bounds.hpp, a wrong workaround for the IBM C++ compiler is applied. This workaround is only valid for IBM C++ on AIX, but wrong for IBM C++ on z/OS. The z/OS compiler has no bug and thus no workaround must be applied (else we get wrong results). So instead of testing against __IBMCPP__ one should also test against __MVS__. This last macro is only defined for IBM C++ on z/OS. I wanted to open a ticket on trac, but this system did always identify my posting as spam. Best regards, Kilian.