On 1/4/11 4:49 PM, %u wrote:
There's still the risk of keeping multiple hashes. Consider:

ulong fun(ulong n) { ... }
alias memoize!fun mfun;

mfun(5); // creates hash ulong[int]
mfun(5u); // creates hash ulong[uint]
mfun('5'); // creates hash ulong[char]


Ohhh I see... so you're basically looking for a compile-time version of overload
resolution, right? Because things seem to be getting complicated very quickly.

It's not that complicated; we'll be able to achieve it somehow. Overloads are compile-time entities so they should be easily inspectable.

(An unrelated side note: I'm new to newsgroups, and I was wondering, what 
program
do people mainly use for communication? I'm using the web interface, but do most
people use Thunderbird or something? Thank you!)

I use Thunderbird with NNTP, which is better than the Web interface.


Andrei

Reply via email to