There are so many other places, so please do it there also, if it is not a cosmetic change.

cc @rafaelfranca

I think this is a cosmetic change.

Hello @divineforest,

Thanks for the patch but we generally do not accept cosmetic patches (see here why). It could be accepted if there was a performance boost but it looks like .keys.each and .each_key are pretty similar regarding performances.

If I'm wrong, feel free to comment so we can reopen this. :-)

@robin850 the benefit is that each_key does not allocate objects and keys allocated new array

hash.keys.each allocates an array of keys; hash.each_key iterates through the keys, without allocating a new array. This is the reason why each_key exists.

👍 for bulk change. So please update in all the places we use keys.each

@rafaelfranca @divineforest Bulk change added in #17099.

@robin850 The reason your benchmark does not show significant performance gains (and my does) is your hash is comprised of symbols, so calling keys only allocates one new object (the array). If the keys in a hash are strings, calling keys results in n + 1 allocations (where n is the number of keys).

Closing in favor of #17099