v1.2.4 breaks compatibility with Node.js 6 about nanoid HOT 8 CLOSED

@AndreyBelym nope, let’s support <6.13.0. If somebody has a problem, we should take care of it.

I am going sleep right now. If you want to make it as quick as possible, make PR with node <6.13.0 to Travis CI and if case with both API.

from nanoid.

Could you please republish v1.2.3 as v1.2.5, and v1.2.4 as v2.0.0 to be more SemVer-friendly, or at least specify the correct engines field in the package.json?

Node.js 6 is still LTS. I care about my users and do not remove active LTS support.

We just need to add a check and use different API for different Node.js versions.

1. Do you want to send PR?
2. Do you want to check tests, why they didn’t fall on Travis CI?

from nanoid.

I found a reason. Node.js 6 backported randomFill https://nodejs.org/en/blog/release/v6.13.0/

from nanoid.

Do we need to fix it in this way? Or we can just ask to update Node.js 6.x to latest minor version?

from nanoid.

I've created this issue because another issue was opened today in the TestCafe repository: DevExpress/testcafe#2905. So after some discussion with the TestCafe team, we've decided to keep support of older Node.js 6.x versions (<6.13.0), because it's easier to e.g. freeze the specific nanoid version than explain to the users that they have to upgrade to the latest minor version.

I can make a PR if you decide to support older LTS versions, but if you don't plan to do it, I can just use v1.2.3 until April 2019 😄.

from nanoid.

Wanted to pipe in and let everyone know that this is also an issue in Node 7.9.0. Apparently crypto.randomFillSync is the problem here as well; it wasn't added until Node 7.10.X.

from nanoid.

Here, docs for node 10.x say it was added 7.10.0
https://nodejs.org/dist/latest-v10.x/docs/api/crypto.html#crypto_crypto_randomfillsync_buffer_offset_size

It is interesting because docs for node 6.x say it was added in 6.13
https://nodejs.org/docs/latest-v6.x/api/crypto.html#crypto_crypto_randomfillsync_buffer_offset_size

from nanoid.

Fixed in 1.2.5 by @AndreyBelym

from nanoid.