Timeline for C++: Smart pointers, Raw pointers, No Pointers?
Current License: CC BY-SA 2.5
4 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Sep 15, 2013 at 13:06 | comment | added | Tim Seguine | I disagree only slightly, specifically with regard to GC. Most of the time, cyclic references are not a problem for reference counted schemes. Generally these cyclic ownership problems come up because people didn't think properly about the ownership of objects. Just because an object needs to point to something, doesn't mean it should own that pointer. The commonly cited example is back pointers in trees, but the parent to the pointer in a tree can safely be a raw pointer without sacrificing safety. | |
| Aug 12, 2010 at 16:38 | comment | added | Simon | Target platform really should influence your design. The hardware that transforms your data can sometimes have large influences on your source-code. The PS3-architecture is a concrete example where you really need to take the hardware into designing your resource and memory-management as well as your renderer. | |
| Aug 11, 2010 at 21:32 | comment | added | Crashworks | Exception handling on the consoles can be a bit dodgy -- the XDK in particular is sort of exception-hostile. | |
| Aug 11, 2010 at 17:42 | history | answered | Jaeger | CC BY-SA 2.5 |