Unless there's a clear benefit I see no need to drop 1.9 for this release.
Granted it's EOL, but there's different concerns w/ Django's policy vs ours.
In contrast - we do have cause to drop 1.8 for this release since it makes one of our issues easier to close off.

In short: it doesn't cost us anything to continue to support 1.9 for now, but it does cost us to continue to support 1.8. That's our guiding line here.

provided that we'll drop it whenever it'll become a pain point what's the point of keeping it ?

what's the point of keeping it ?

I'm notthat fussed either way, but...

In order to better support teams who started their project in the first half of 2016, who'd like to stay up to date with the latest version of REST framework, and don't have any pressing reasons to upgrade from Django 1.9. (Despite the formal EOL)

@xordoquy@carltongibson - Happy to defer to your judgement here, tho. Go for it! 😄

OK, cool.

Just for anyone reading this, my views are:

1. Formal EOLis a pressing reason to upgrade. This is simply from the security aspect. It's really not OK to be using EOL software.
2. The non-LTS versions are explicitly for people who are in a position to keep up. Others should use LTS.

For both these reasons I'm for keeping the matrix rolling forwards — otherwise we send the wrong message to users.

But also, it does make our job easier:#5458

One minor reason to drop 1.9 is that it reduces the total build time/saves trees.