On Tue, Mar 27, 2007 at 04:32:13PM +0200, Abdelrazak Younes wrote: > Andre Poenitz wrote: > >On Tue, Mar 27, 2007 at 02:00:13PM +0200, Abdelrazak Younes wrote: > >>One tiny comment: maybe you could use a boost::scoped_ptr instead of a > >>raw pointer for undo::pars? > > > >*shrug* I don't see the benefit. I put ASSERTs all over the place > >to make sure we walk on the safe side. > > That's the point really: less loc and be on the safe side if someone > else less paranoid than you use the pointer ;-)
Well, I _do_ know how bald pointers work. I also know how I can shoot myself in the knee when using them and I also know how to avoid that. Also, it is in ISO 14882. I do not know what boost::scoped_ptr might be doing (today, tomorrow, ...) behind my back. As far as I can tell parts of boost differ widely in quality. Using scoped_ptr certainly costs more cycles. This might not be relevant in this particular case but I am a fan of that "don't pay for something you don't need" thingy. Andre'
