Pavel Vozenilek <pavel_vozenilek <at> hotmail.com> writes:
Hello Joaquín,
Yes, I think there's no problem at all in implementing this extension. Consider it done (if I run into trouble I'll let you know.)
Maybe the feature should wait a bit.
I know to sound conservative but: - I know no container library with this feature - it does something at very core level
It may be better to discuss it before adding.
I personally do not see anything bad with it but I'm not labguage lawyer.
I'll happily do whatever the people agree on. Personally, I dont't see any core problem in the extension. The main reason I deem the change harmless is because we are replacing one function with one function: if we had to provide different overloads for say operator==, function template ordering would play a role, which is more dangerous (some compilers do not support this right.) My conclusion is then that the change cannot posibly break anything --but as I said before, I'll abide the mailing list's verdict. Joaquín M López Muñoz Telefónica, Investigación y Desarrollo