Memoize is an interesting proof of concept, but it is so easy to do simple caching like this that I find no use for it. If I had a candidate for memoization that had many different patterns of parameters or possible return contexts, I might consider it. But even then, you can do a whole lot of stuff in the time taken by Memoize's overhead so I'd want to see benchmarks to prove it was actually of use.
Use PadWalker to access %cache outside the sub. (I'm guessing it should work on 5.10's state variables, but don't know it for a fact.)
In reply to Re: How to test caching?
by ysth
in thread How to test caching?
by JavaFan
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |