r-base installation fails on Ubuntu 14.04

He was suggesting to upgrade to 15.10 if you needed the newer fortran
for some reason (or other newer things only in backports or not even there).

I also mostly run LTS versions of Ubuntu, especially in my research
group where I manage our cloud/servers. Which also matches the computer
cluster available to me. That said I also stick to the stock packages
for all the basic underlying libraries like fortran. If for some reason
I need super new stuff to run a specific tool/analysis that's when I
spin up a custom VM/Docker to create such an environment.

In this case I can understand, 14.04 came from the hardware vendor, so
it makes sense to stick with that for now given the support from the
company for drivers (hopefully they will also support 16.04). In which
case, yes roll back your backported fortran to use the version Ubuntu
originally supplied.

Thanks,
Alex

-------- Forwarded Message --------
Subject: 	Re: [R-sig-Debian] r-base installation fails on Ubuntu 14.04
Date: 	Wed, 23 Mar 2016 13:07:44 -0700
From: 	Barnet Wagman <bdwgen at gmail.com>
To: 	Dirk Eddelbuettel <edd at debian.org>

Lastly, and please don't take this the wrong way: I think I am helping way
more people like you who for one reason or other insist on older / frozen
system like 14.04 but then desire newer software.  Simply running _current_
Ubuntu and upgrading every six months is IMHO much easier.

Are suggesting upgrading to Ubuntu 15 or just running apt-get upgrade or
apt-get dist-upgrade (which I've done).  Ubuntu 14.04 is the latest LTS
version, which is probably why a lot of people are not going to 15.


PS Inadvertently sent this to Dirk instead of list.  Sorry.

| He was suggesting to upgrade to 15.10 if you needed the newer fortran
| for some reason (or other newer things only in backports or not even there).

Yup.

Which from 14.04 requires upgrades to 14.10, then 15.04, then 15.10.  In
short may _right now_ also just wait for 16.04 LTS.

| I also mostly run LTS versions of Ubuntu, especially in my research
| group where I manage our cloud/servers. Which also matches the computer
| cluster available to me. That said I also stick to the stock packages
| for all the basic underlying libraries like fortran. If for some reason
| I need super new stuff to run a specific tool/analysis that's when I
| spin up a custom VM/Docker to create such an environment.

If you must keep on a version it can also help to __build you own backports__
which by being built locally will match your packages.  Explaining how to do
that is beyond the scope of this thread which failed over much simpler things.

Dirk

http://dirk.eddelbuettel.com | @eddelbuettel | edd at debian.org