On 2009-11-30 at 16:11:48 [+0100], Stefano Ceccherini <stefano.ceccherini@xxxxxxxxx> wrote: > 2009/11/30 Ingo Weinhold <ingo_weinhold@xxxxxx>: > > > Yep. There's quite a bit of static libbe initialization going on, which is > > wasted time for programs that don't use the respective features. Like the > > build tools that link against libbe. That won't cut the build times > > dramatically, but it's easily done at least. > > > > Since you're doing this, watch out for the problem described in ticket > #505 (which I lazily fixed by moving the computationally-intensive > stuff to libbe initialization). Thanks for the hint. ATM the palette initialization happens when the BApplication is constructed, not at libbe initialization time (only a lock is created then). It would be nice to postpone that even further, though. CU, Ingo