OK, good enough for me, although I don’t like enticing people to use APIs that are slow when there’s a way that is 5 times faster. Maybe we should warn folks about this in the asdoc?
-Alex On 5/19/15, 2:25 PM, "Gareth Smith" <garethdanielsm...@gmail.com> wrote: >I have done some very unscientific tests: for each kind of list (Array, >old ArrayList, new ArrayList, new ArrayList - using sugar) the test code >creates a million-item array and then increments each value that it >contains. > >This is how long each test takes: > >Array: 0.1 sec >old ArrayList (get/set Item): 0.8 sec >new ArrayList (get/set Item): 0.9 sec >new ArrayList ([]): 5.6 sec > >In summary, indexing sugar is really slow, but get/set item don't seem >to get much slower just by implementing Proxy. > >Gareth > >On 18/05/15 06:56, Alex Harui wrote: >> >> On 5/17/15, 9:22 AM, "Gareth Smith" <garethdanielsm...@gmail.com> wrote: >> >>> I don't know about any performance impact. >> Can you run a some of quick tests? Iterating on 1000 items via for-each >> on an Array, via a loop on ArrayList.getItemAt() before your Proxy >>change >> and using [] and for-each on ArrayList after your new Proxy change? >> >> And maybe also try 5000 and 10000 items if you have time. >> >> Thanks, >> -Alex >> >
