LinuxQuestions.org - Recipe for installing a recent RStudio on current64

- Slackware (https://www.linuxquestions.org/questions/slackware-14/)

- - Recipe for installing a recent RStudio on current64 (https://www.linuxquestions.org/questions/slackware-14/recipe-for-installing-a-recent-rstudio-on-current64-4175692102/)

Recipe for installing a recent RStudio on current64

Hello All

I'm testing current on an old laptop. I skipped the kde package set, apart from that it is a full 64 bit install, updated to today. I'm using xfce4.

R is a command line based statistical language that is quite popular. There is a slackbuild for it on slackbuilds.org and it is a simple compilation with no external dependencies.

RStudio is an IDE for R which is often referred to in books about R. RStudio distribute binary builds for .rpm and for .deb. The source code is available but it is a complex compilation with a lot of dependencies. There is a slackbuild for version 1.1.4 of RStudio that re-packages the .deb based binary. RStudio 1.1.4 is the last version (I think) that supports 32 bit, the newer ones are 64 bit only. I decided to try installing the most recent version of RStudio, 1.4.x on current.

After some head scratching and googling, I found that you need to

install postressql from slackbuilds. I just used the 14.2 slackbuild with no issues
download and convert the compat-openssl10-1.0.2o-3.el8.x86_64.tgz package from the CentOS 8 repositories and install the resulting Slack package
download and convert the RStudio .rpm binary package and install the resulting Slack package

Here 'convert' means using rpm2tgz. I'm familiar with that script so that is what I used.

The T61p Thinkpad has nvidia graphics and I have not installed the nvidia drivers, so I had to add the line below to RStudio's desktop.ini file to disable accelerated graphics for just RStudio.

Code:

ash-5.1$ cat /home/keith/.config/RStudio/desktop.ini

[General]

...

desktop.renderingEngine=software

...

I'm wondering if this recipe looks sort of sensible and/or there is an obviously easier way. It works fine... at the moment :foot:

After building any new version of R, my first three commands are:

Code:

chooseCRANmirror()

install.packages("Rcmdr", dep=T)

library(Rcmdr)

That installs and runs R Commander, which I have found to be a very useable interface to R. IMHO, Rstudio is overkill.

Quote:

Originally Posted by allend (Post 6230831)

That installs and runs R Commander, which I have found to be a very useable interface to R.

Thanks for your reply.

Installs using those commands as user on current fine, and yes, it is what I would call a graphical interface to an R session. I also like how it compiles in the user directory - takes a bit of time on a core duo.

At present, the book I'm working through is based around RStudio, which, as you say, is an IDE with bells and whistles. I quite like the tab-completion with links to help &c.

Takes even longer on an Atom N270, although these days I set up icecream to a beefier machine before building. The bottleneck is still the linking, but all the compiling ends up being passed to the beefier machine.
I do not miss tab completion as most of the time I end up needing to replicate small stanzas with changes where tab completion would not help.
My R scripts are often sprinkled with ?<something> lines to bring up the documentation. I also seemed to have trained Google to quickly point me to relevant R guides.

Glad you are happy. Good luck!