search

brodieG / diffobj

Compare R Objects with a Diff
229 stars 12 forks source link

Support R 3.1? #123

Closed krlmlr closed 6 years ago

krlmlr commented 6 years ago

Tried to set up a testing environment for R 3.1 for fansi, but noticed that diffobj requires R 3.2. What's the reason?

brodieG commented 6 years ago

No specific reason other than 3.2 felt far enough back and I wasn't testing that far back. There might be something that breaks in 3.1, but probably not.

Now that you mention this, this might be the reason the 3.1 test broke wrt to unitizer, since it uses diffobj. What I don't understand is how it could have ever worked in the first place, as it did when I tested it for the original release. This is something I'll need to look into more.

brodieG commented 6 years ago

Ok, I figured out what happened. Since releasing fansi, I added the Depends >= 3.2 to diffobj, thinking that was a smart thing to do given that I do not test diffobj for <= 3.2... I'll need to fix that too.

krlmlr commented 6 years ago

Thanks! The tidyverse packages aim at supporting R >= 3.1, it's a baseline not too far back in the past.

brodieG commented 6 years ago

Yes, I just failed to remember that fansi has a suggests dependency on diffobj, and as such diffobj also needs to be R3.1 like fansi. It was a careless mistake. I'll address this first and get it on CRAN, and then the fansi issues (otherwise tests will keep failing on 3.1).

brodieG commented 6 years ago

Fyi, working on this now. There are a couple of potential issues I need to figure out. This will be a blocking issue for all the fansi issues that need to be resolved. I hope to get this done in the next couple of days.

brodieG commented 6 years ago

This is now fixed on the development branch. Took longer than I expected as there was a surprising (to me) amount of code that relied on > 3.2.

krlmlr commented 6 years ago

Thanks for taking the effort!

brodieG commented 6 years ago

Not at all! When I agreed to make fansi 3.1 compatible, I implicitly committed myself to do the same for diffobj, even though I didn't quite realize it at the time.