promisify about yaku HOT 4 CLOSED

Good point, I will optimize the performance later. Besides, you can try my another lib nofs.

from yaku.

Looks nice, in bluebird I just Promise.promifiyAll(require("fs")) and have a (close to) 0 overhead wrapper over fs, if I want to map/reduce etc I just use bluebird's utilities.

Your code, uses Yaku's promisification in the test - I expect it to perform an order of magnitude slower if the OS caches the read files. It creates a closure and a bunch of other stuff whenever it reads a file :)

When you optimize promisify I expect it to perform a lot better :)

P.S I wonder - why do you have a readJsonSync function there? That's practially just a require

P.S 2 - why don't you chain in the tests https://github.com/ysmood/nofs/blob/1a0fe800548a8d58862b276c39fb7e960fe0fb41/test/basic.coffee#L51-L53 ?

P.S 3 - this is just Array.prototype.every https://github.com/ysmood/nofs/blob/master/src/utils.coffee#L32-L37

P.S 4 - https://github.com/ysmood/nofs/blob/master/src/utils.coffee#L32-L37 and this is Array.prototype.some

from yaku.

P.S I wonder - why do you have a readJsonSync function there? That's practially just a require

Because require will cache the file, readJson won't waste your memory.

P.S 2 - why don't you chain in the tests

Coffeescript will auto chain them, or I mistook your meaning.

P.S 3 - this is just Array.prototype.every
Array.prototype.some

Old node don't supports them.

About your performance concern, I agree, but my 1 year experience told me it won't be the bottleneck of a large server side project, the IO itself will alway be the bottleneck. So we don't have to waste too much time on it. A better way is to push node support it natively, such as the browsers provide promise based fetch IO, all the file system operation should also support it internally.

from yaku.

b336a4d#diff-1f0dd9d666794c01ded659349cb2512eR20

The promisify of Yaku has much less overhead now, the inline codes made the trick.

from yaku.