Maintainers

[kushkella]

Hello -

I'm soon going to deploy company's RedShift data warehouse and this gem is a major dependency both for querying and running database migrations. I identified few issues that I came across and fixed those. But I see there are several open PRs and issues. If you need help maintaining the gem, let me know. I can help out and also add tests.

If there is anyone else who want to maintain the gem, please comment so @aamine can know how eager we are to see new versions of this gem :)

Thanks a lot!

[CallumD]

Hi @kushkella are you maintaining a fork of this gem?

[lorgio]

@aamine hasn't done much on the keeping up to to date with issues/errors/pull requests. i've forked off and done my own updates when i see errors. I've put in PR's but they seem to go untouched. for the most part the gem works, but does have it's own issues..

[CallumD]

@lorgio perhaps we can sync up with @kushkella to have an actively maintained fork of this gem?

[kushkella]

No I'm not. I've added some helper code to my project to help with unit testing, but I haven't created a public fork. I submitted a PR to fix issues in this project. If we team up, we can probably use the general architecture of postgres adapter but build it from scratch for redshift.

Following are some pain points I faced while using this project/gem:

1. The adapter still uses a lot of Postgres related operations that aren't supported by Redshift. And it really tries to do all the Postgres stuff in Redshift.
2. There is no way to use this adapter in test mode. Personally I've added some code that allows me to
   use a Postgres database as a replacement of redshift, and it works well.
3. There is no test coverage which is why it is risky to use in production.

Thanks!

[kushkella]

And @lorgio I apologize for not being able to merge your PR in my fork given I wasn't maintaining it.

[CallumD]

I think it depends on how much time people are prepared to commit to this. I would like to see a production viable redshift active record adaptor but I don't know that I have the knowledge or the time to have a significant input on a brand new adaptor.

I would be happy to help contribute however.

[cobbr2]

We've maintained our own fork for a while, but only added our own local (and not particularly wonderful) patches. I think our folks would be happy to be more active, both on reviews and issue work (at least for issues relative to our use cases), but unlikely to take the lead.

As far as where to start to get to higher quality, we've been happy enough with patch-until-fixed; I'm not sure what "build from scratch" would mean in the context of an AR adapter.