1 2History of Major Changes to TestFloat, up to Release 2a 3 4John R. Hauser 51998 December 17 6 7 8The TestFloat releases parallel those of SoftFloat, on which TestFloat is 9based. Each TestFloat release also incorporates all bug fixes from the 10corresponding release of SoftFloat. 11 12- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 13Release 2a (1998 December) 14 15-- Added support for testing conversions between floating-point and 64-bit 16 integers. 17 18-- Improved the makefiles. 19 20- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 21Release 2 (1997 June) 22 23-- Integrated the generation of test cases and the checking of system 24 results into a single program. (Before they were separate programs, 25 normally joined by explicit command-line pipes.) 26 27-- Improved the sequence of test cases. 28 29-- Added support for testing extended double precision and quadruple 30 precision. 31 32-- Made program output more readable, and added new command arguments. 33 34-- Reduced dependence on the quality of the standard `random' function for 35 generating test cases. (Previously naively expected `random' to be able 36 to generate good random bits for the entire machine word width.) 37 38-- Created `testsoftfloat', with its own simpler complete software floating- 39 point (``slowfloat'') for comparison purposes. 40 41-- Made some changes to the source file structure, including renaming 42 `environment.h' to `milieu.h' (to avoid confusion with environment 43 variables). 44 45- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 46Release 1a (1996 July) 47 48-- Added the `-tininessbefore' and `-tininessafter' options to control 49 whether tininess should be detected before or after rounding. 50 51- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 52Release 1 (1996 July) 53 54-- Original release. 55 56- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 57 58