Rosa Gutierrez

@byroot thank you! Already did: - https://github.com/rails/rails/issues/51388 - https://github.com/rails/rails/issues/51392 Both have repro scripts.

I wanted to add another case that we ran into in Basecamp, and it made me more convinced that this change is too confusing, introducing some really sneaky bugs and...

@byroot, I just realised that: > To solve your problem short term, I believe that setting `inverse_of: false` on that impacted relation should restore the old behavior for you. started...

Ok, I got a reproduction script for a case where `inverse_of: false` breaks previous behaviour and doesn't work as a way to opt out of `automatically_invert_plural_associations = true`. Adding it...

It seems this also breaks URL helpers in helper tests that extend `ActionView::TestCase`, as route drawing hasn't happened and any URL helper is undefined. Not sure if it's intended. For...

@rafaelfranca thanks! I haven't been able to create a minimal reproduction script, though 😕 Do you still want an issue without that?

@tomafro [this](https://github.com/minad/mimemagic/pull/52) might be relevant for `.xlsx, .docx, .pptx`. I think we are failing right now to correctly map all those file types if they were generated with something else...

Hey @bleonard! This week I've been working on a `CookieStore` for Makara, since right now we are using it without stickiness because we don't set any kind of backend cache...

Thanks for your reply and your notes @bleonard! You are correct, I was talking about two different things in my comment, and I didn't explain myself very well for the...

Oh! A concern I had with the above solution was the possibility of exceeding the cookie size limit for browsers (assuming only modern browsers that support something like 4096 bytes...