options.redis.opts.createClient no longer works?

[ianstormtaylor]

I've been trying to track down tests failing in super weird ways for many hours now, and it seems like in 3.x the options.redis.opts.createClient function is no longer called? I used to have it setup like this:

const client = new Redis(REDIS_URL)
const subscriber = new Redis(REDIS_URL)

function createQueue(name) {
  const queue = new Queue(name, {
    redis: {
      opts: {
        createClient(type) {
          switch (type) {
            case 'client': return client
            case 'subscriber': return subscriber
            default: return new Redis(REDIS_URL)
          }
        }
      }
    }
  })
}

Which is still how it's recommended in the PATTERNS.md document. And the MIGRATION.md document doesn't really mention what has changed.

Is there another way to do this now that's not documented?

[ianstormtaylor]

For anyone who stumbles on this, apparently createClient has been moved to a top-level option:

const client = new Redis(REDIS_URL)
const subscriber = new Redis(REDIS_URL)

function createQueue(name) {
  const queue = new Queue(name, {
    createClient(type) {
      switch (type) {
        case 'client': return client
        case 'subscriber': return subscriber
        default: return new Redis(REDIS_URL)
      }
    }
  })
}

[ritter]

For what it's worth ... you could update the documentation and submit a PR. I'm sure contributions like that are quite welcomed.

Was also mentioned in #545

[manast]

I have updated it now.

[ianstormtaylor]

@manast thank you! I think in the future, the most helpful thing for me as a consumer of this library would be if the CHANGELOG.md was a bit more descriptive, and ensured to contain every possible breaking change in a very obvious way. Ideally they'd be bolded or something else to quickly scan. (Even if that meant not mentioning the new additions, since those are lower priority if there's not enough time.)

@ritter thanks, I definitely realize that—here's a past PR I made to the docs. In this case though I think it's pretty hard to ask people to contribute docs for undocumented breaking API changes they don't even know about when they go to upgrade.