Previously, converting a string into a decimal floating point number greater than zero and less than one caused an error. During this conversion, the first decimal digit disappeared; consequently all following computations were done with wrong numbers. This bug has been fixed so that the conversion into the decimal floating point number now works as expected. (BZ#625495)

Under some circumstances, libdfp encountered an issue while converting a value from a string into a decimal floating point number with the conversion command "strtod32". The "strtod64" and "strtod128" commands, which are also included in libdfp, did work correctly. The problem has been resolved so that the conversion now proceeds properly. (BZ#628670)

Previously, there were several testsuite failures encountered when building the libdfp packages for the IBM S/390 architecture. These testsuite failures have been corrected with this update and thus no longer occur. (BZ#673222)