Update R to 3.4.1

[kylehamilton]

I'm trying to use the weightr package with a Jamovi module but the current version of R that Jamovi uses is 3.3.1. Weightr requires at least R 3.3.3.

[jonathon-love]

yeah, righto. big R version updates are a bit painful, but we'll see what we can do.

jonathon

[jonathon-love]

OK, i think i have this working. will try and have this out in the next couple of days.

[jonathon-love]

(other platforms will probably lag a bit ... or not keep up at all -- xenial/mint only ship with 3.2 for example).

[kylehamilton]

Sounds good and thanks a lot I didn't know it would be such a pain I'll take a look later today! Since p-curve doesn't have a useful R package on CRAN the only modern publication bias option was weightr.

[dropmann]

This was released in 0.7.5.4 :) can probably be closed.

[kylehamilton]

Thanks!

[jonathon-love]

so i've only updated to R 3.4.1 under windows so far, macOS is still on 3.3.0.

so depending on what your package uses, you should tell people it will be windows only for the time being.

i'll update the macOS version in the next while, but the ubuntu LTS's will be stuck on 3.2.3

[odeleongt]

I've not found it, and maybe it is out there, but have you considered documenting the build process? I wouldn't mind having to build jamovi for different R versions for package testing / module dev purposes.

[jonathon-love]

hi @odeleongt,

we have some documentation here

in order to use a different R version, you'll only need to rebuild the engine component. you can change the makefile to suit:

https://github.com/jamovi/jamovi/blob/master/engine/Makefile

finally, you'll need to update the env.conf file to the new location:

https://dev.jamovi.org/info_project-structure.html#bin

having said all that, for testing purposes, it's probably easier to simply build jamovi modules into R packages, and test them in R.

let me know if you have any more Qs.

[odeleongt]

Thanks @jonathon-love, will do.

[deepankardatta]

Is there any progress on updating the Mac version of jamovi?

Big fan of the concept of your software, but unfortunately the CRAN package that I want to link to is based on R 3.4.0, and Mac Jamovi is 3.3.0 from what I can see from the jmvtools::install() output.

[jonathon-love]

hi,

yes, it isn't too much work to transition the mac version to R 3.4, the issue is that R 3.4 isn't compatible with macOS 10.10 or 10.9. these apparently constitute ~18% of mac users

we'll definitely make the jump at some point, but have been holding off for now.

what packages are you wanting to use? it's often possible to just hack the source packages and remove the >= 3.4 requirement, and have them still work.

give that a shot, and if that doesn't work, i guess we'll have to have a serious think about updating the R.

cheers

jonathon

[deepankardatta]

Hi

Thanks for the reply, and interesting fact about 3.4 compatibility.

The package I wanted to use was something I wrote myself called blandr (can see it on my github, or https://cran.r-project.org/package=blandr). It's not that advanced a package, and am pretty sure it really doesn't need 3.4. I can't quite recall why the final package needed 3.4, but I think it was something to do with CRAN...

Anyway I'll see if I can use "install_github" in the library to use a branched version which is 3.3 compatible, and then see if I can modify the next CRAN release.

[jonathon-love]

if it's your R package, it might be easier to just turn it into a jamovi module. that way it can be built into both an R package, and a jamovi module, and the R syntax generated by jamovi will make sense.

[jonathon-love]

all 'current' versions use R 3.6.1 now.