Simplify seq.nanoduration for devel bit64 - #152

Note that it crashes in CI presumably because CI at current only knows CRAN bit64.

Note that it crashes in CI presumably because CI at current only knows CRAN bit64.

Good point. I edited so that the new approach is only done conditionally.

Given that it is 'breaking behaviour' I would prefer conditional use, ideally in the code rather than the tests. Is that possible?

If it really really is hard requirement we can put it into DESCRIPTION would I would prefer to cast a wider, easier-on-users net. Sometimes we can, sometimes we can't.

Agreed. It's branched now so we are green here with pre- and post-4.8.0 {bit64}.

thanks for insisting on higher-quality code!

I am in between three things but I will try to check here too. I was a little puzzled I had breakage so I want to take another look. I think when I pulled last bit64 still say 4.7.99. I guess I need to pull too? You test used '4.8.0'.

You test used '4.8.0'.

Of course now I realize I numbered the release incorrectly 🙈 (it should be 4.7.0)

Per Uwe 4.8.0 will be released from purgatory momentarily, so I would just wait for that to be official.

it should be 4.7.0

If you email them right now you can probably get 4.8.0 nixed away and re-upload as 4.7.0.

Locally, and with 4.8.0 I still get 50 fails in test_nanoival.R

If you email them right now you can probably get 4.8.0 nixed away and re-upload as 4.7.0.

(for completeness, yes, I got there in time, but I decided against it because my dev version had been 4.7.99 for some time, and I think it's more preferable for the CRAN release to come after dev than it is to have strictly correct increment of the CRAN version number, i.e., the real mistake/original sin was the wrong dev version number for ~1 year. thanks for the encouragement to get it fixed though!)