Stephan Schiffel wrote: > Hi all, > > I think I found a bug in ECLiPSe (6.0 #69): > > %%% file: eclipse6_bug.pl > :- local store(options). > set_option(Name, Value) :- > true, > store_set(options, Name, Value). > %%%% > > Compiling above file fails with the following output: > > [eclipse 1]: ["eclipse6_bug"]. > ... > Check failed: 0 == 0, nonvar(_76001) > Aborting execution ... > Abort > > The code works as expected if "true" in the first line of set_option/2 is > removed or replaced by anything else. Thanks, fixed for 6.0#72. As you know, ECLiPSe 6.0 has a completely new compiler, so some teething problems were expected, although it has done pretty well so far. It's still amazing how simple failure cases like this can go unnoticed for a couple of months, when they would normally not be written by a user in such a way. There is something to be said for testing a compiler with randomly generated input! -- JoachimReceived on Sun Feb 08 2009 - 03:50:19 CET
This archive was generated by hypermail 2.2.0 : Thu Feb 02 2012 - 02:31:58 CET