[R-pkg-devel] "try" malfunctions on Ubuntu Linux 16.04 LTS, R-release, GCC

[profjohn]

Looks like Spencer's issue solved in another post.

And this isn't a criticism of Dirk's contribution, just a partial
explanation of some of the noise that occurs.

I use Linux Mint -- family members are still recovering from Win XP.
But LM 18.x is based on Ubuntu 16.04. So I found recently that the
CRAN check_packages_in_dir() (if I remember the syntax) and Gabor C's
package revdepcheck didn't work. Latter reported as an issue.

On two of my machines I have LM 19.x. However, I'm travelling, and so
I tried upgrading one of my laptops with mintupgrade. Bad idea -- got a
very weird lockup situation -- I could boot to a login screen, but that
would accept the password and return to the login screen. Had to do a
bare metal reinstall. Painful, but revdepcheck now works and revealed a
small but vital typo in one line of my package. Sigh.

While there are huge benefits to upgrading, there is pain. It's that
pain that keeps people on OS versions that are older but still in the
LTS period.

Best, JN


On 2/3/20 12:27 AM, Dirk Eddelbuettel wrote:
> 
> On 2 February 2020 at 21:47, Ben Bolker wrote:
> |   Maybe an Rhub glitch.  I just ran R CMD check with a recent r-devel
> | (2019-12-03 r77509) on Ubuntu 16.04 with no problems.
> 
> Is there a reason that keeps _both_ of you on Ubuntu 16.04 which has been
> replaced _nearly two years ago_ by the subsequent LTS release 18.04 ?
> 
> Michael works hard to provide _an awesome_ experience using the current LTS
> release_ 18.04 that I am really hard-pressed to see why 16.04 would matter.
> 
> Dirk
>