> After I make a release branch, I noticed there are some test case > failures for the Data class that need to be looked into, and a few > other > things.
If you mean the Date class, there are some failures on Debian Etch due to the compiler outputting "0" for ostream& << showpos << 0, instead of "+0", as the C++ Standard dictates (bug https://savannah.gnu.org/bugs/?24752). There are also some on i586 gcc 4.3 (Debian Lenny and Mandriva 2009 at least) due to compiler floating point differences. The problem is that the Date class uses a double internally, and C++ allows some weirdness when optimizing floating point operations. We test some large-number Date representations in the testsuite; some compilers do the same as the Adobe player, some don't. Both outputs are Standard-compliant. They can possibly be fixed (to make them compatible) using a gcc compiler flag, or perhaps by trying to enforce storage of an intermediate stage in the calculation, but both of these options would reduce performance for the sake of fixing fairly insane cases. bwy -- Yes, YouTube does work in Gnash http://www.gnu.org/software/gnash/ Benjamin Wolsey, Software Developer - http://benjaminwolsey.de
signature.asc
Description: Dies ist ein digital signierter Nachrichtenteil
_______________________________________________ Gnash-dev mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnash-dev
