RE: [AD] Namespace again (was Problems with gcc 3.1) |
[ Thread Index | Date Index | More lists.liballeg.org/allegro-developers Archives ]
> - newcomers will find since the beginning a clean and robust API,
> without any namespace conflicts, and they'll just love Allegro =)
This ought to go with a clean and robust *implementation*.
However, this ain't the case. There are plenty of places where code
doesn't care about buffer overflows. Plenty of "char tmp[64]", uses
of sprintf (yes, I know snprintf is not ANSI, but it is (or another
named equivalent) widely present), still too many bugs in the unicode
stuff (which wasn't ever really tested due to no people actually
using it).
I think it is irrealistic to believe that there will be only one
WIP between the feature freeze and the release. Testing should be
greater, and focused on the parts that aren't get tested normally
(eg unicode...)
A good way to do that would maybe to release an official beta version
instead of an official version. Make people know that this is beta,
that the interface is stable, but that it may have some bugs. This
would get other people to test, and they'd likely test other stuff,
or differently, allowing more bugs to surface.
--
Vincent Penquerc'h
Mail converted by MHonArc 2.6.19+ | http://listengine.tuxfamily.org/ |