madmats@elcgl.epfl.ch (03/27/91)
In article <jls.669955177@rutabaga>, jls@rutabaga.Rational.COM (Jim Showalter) writes: >[stuff deleted] >> The increased focus on usage and the prohibition of continued >> growth is expected to provide vendors with more time to concen- >> trate on meeting users special requirements for optimization and >> high quality tools. > > I would rephrase this to read "The prohibition of continued growth > [of the tests] is expected to provide vendors with a shortcut to > achieving validation at the expense of the Ada user community". Jim, I like your reply to the new ACVC policy. One must realize, however, that the current policy is somewhat inappropriate: still now, 8 years after the Ada standard has been published, many compilers lack quality in compiling real code. Language users probably have more need for a compiler that handles correct constructs right and occasionally compiles an incorrect construct than the reverse. The current ACVC seems to test just the reverse and this causes many problems. For instance, we (Software Engineering lab, Swiss federal institute of technology), have had problems with compiling our reusable components developped using nested generics from the beginning, and the situation has not improved very much, to the point that I wonder if most Ada 83 compilers will be correct before Ada 9X. While I am at the subject, I have extracted a sample (12000 LOC) of our test programs that I have made completely implementation independant and organized into a test suite with input and expected output files. I consider these tests to be 'normal' uses of Ada's generic facility, but many available complilers don't compile them. The test suite is available through anonymous ftp on elcgl.epfl.ch (128.178.76.5) in directory adabugs (beware: this is a VMS machine). Any comments are welcome. I intend to 'organize' this test suite a little more, publishing its location and possibly adding more tests in the future. Unfortunately, I have no better means than human reading for proving the correctness of the code in the test suite, as there is no formal specification for Ada :-) Mats Weber Swiss Federal Institute of Technology EPFL DI LGL 1015 Lausanne Switzerland E-mail : madmats@elcgl.epfl.ch phone : +41 21 693 52 92 fax : +41 21 693 39 09