Refine _.pick

[jridgewell]

Not sure how this’ll do performance wise. I’ll write a jsperf tomorrow.

[megawac]

Any advantage to this refactor? Did _.pick({constructor: 1}, 'constructor') not work or something? (thought 173a7fa fixed this)

[jridgewell]

It was a lot of duplicated code between the two branches to support an inline key in obj.

173a7fa fixed issues with adding new keys as you're iterating, but it created linting errors (duplicate var declarations).

---

I think this wraps up all the removals changes from 1.8 to 1.8.1. I can't spell anymore, time for bed. 😴

[megawac]

I tried a similar refactor before for reference #1639. Had heavy perf hits for the keys case

[jridgewell]

That's probably because of Line 903. The only perf hit this should take is invoking (and this is sure to be inlined) in the _.pick(obj, 'key', 'key2') case.

var iteratee = function(value, key, object) {
  return key in object;
};

[megawac]

Can you post benchmarks when you get a chance
On Feb 20, 2015 1:06 AM, "Jeremy Ashkenas" notifications@github.com wrote:

Merged #2065 #2065.

—
Reply to this email directly or view it on GitHub
#2065 (comment).

[jridgewell]

jsperf.

Unfortunately, it's a bit slower for the _.pick(obj, 'key') case. I was able to improve performance a bit in an opt branch.

[jashkenas]

That's a weird optimization, no?

[jridgewell]

Are you talking about #2068?