Andrew Harper

This seems like a plausible bug to me. Is it possible to write a new TinyTDS test case to verify the expected behavior?

This seems like it's more related to the AR SQL Server adapter, so probably best to ask over there: https://github.com/rails-sqlserver/activerecord-sqlserver-adapter/issues

TinyTDS can't really tell you anything that FreeTDS isn't already saying; ie, it can't connect for some reason. The diagnostics are up to you. Here's some ideas to get you...

Looks like the default changed [here](https://github.com/rails-sqlserver/tiny_tds/commit/d5d86c45afeb36daf2b0bf53c219906cf5f5d34c) with the switch to FreeTDS 1.0. Can you open a PR to update the readme?

Does `:use_utf16` already work like you want? https://github.com/rails-sqlserver/tiny_tds/blob/master/lib/tiny_tds/client.rb#L52 > all settings in DATABASE_URL, such as "?azure=false" should return "expected" config result of boolean false, consistent with ActiveRecord conversion from database...

I'm somewhat skeptical that that code should live here in the tiny_tds project. To a reasonable degree, this project should be web framework agnostic and follow normal gem paradigms. Treating...

> unless we want to train the SQLServer adapter how to recognize those new config keys Yeah, I'm thinking that is possible the preferable way to do it. That seems...

I've speculated about this exact concept myself. I think it's interesting. But I agree with @wpolicarpo, it'd probably just need to be a new from-scratch project. I think that designing...

Only problem I can think of with that approach is that defaults will now be applied to a copy of the input parameter, rather than mutating it. That means any...

I believe this will be addressed by https://github.com/rails-sqlserver/tiny_tds/pull/475