[R-sig-Debian] [FORGED] r-base is already the newest version (3.5.2-1bionic)

Rolf Turner r@turner @end|ng |rom @uck|@nd@@c@nz
Wed Jan 30 03:18:13 CET 2019


On 1/30/19 2:03 PM, Dirk Eddelbuettel wrote:
> 
> On 30 January 2019 at 12:33, Rolf Turner wrote:
> | On 1/30/19 11:53 AM, Dirk Eddelbuettel wrote:
> | >
> | > Rolf,
> | >
> | > I think it may help to read-up on dpkg and apt. Instead of 'whereis' do
> | >
> | >    dpkg -l r-base-core
> 
> And I also meant  'dpkg -L r-base-core'.  Both -l and -L do useful (but
> different) things.

Well I'm glad you made that error.  Regardless of the fact that 
miss-interpreted the output, that from "dpkg -l" set me going on a 
procedure that in the end worked.  The output from "dpkg -L" would 
simply have bewildered me.

>   
> | Now *that* was a revealing suggestion!  I did that and got:
> |
> | > Desired=Unknown/Install/Remove/Purge/Hold
> | > | Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend
> | > |/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad)
> | > ||/ Name             Version       Architecture  Description
> | > +++-================-=============-=============-======================================
> | > ii  r-base-core      3.5.2-1bionic amd64         GNU R core of statistical computation
> |
> | So indeed the installation had not actually been done.
> 
> Why do you say that?  That is the __normal__ display after a successful installation.

I interpreted the "Status=Not/Inst ...." to mean that the package was 
not installed.  Which is consistent with what was actually the case.

>   
> | The nature of the message prompted me to try
> |
> |      sudo apt purge r-base-core
> |
> | which ran and said that it was removing a whole lot of stuff.
> 
> Why?

I figured that this would get rid if the not installed/partially 
installed/buggered-up-installed traces of r-base-core, and let me try 
again.  In the event this seemed to work.

>   
> | I then tried
> |
> |       sudo apt-get install r-base-core
> |
> | which ran and gave a whole lot more output than was previously produced
> | when I ran "sudo apt-get install r-base".  (Note: I had been typing
> | *r-base* and NOT *r-base-core*.)  It indicated that it was doing lots of
> | stuff that looked promising in respect of actually *installing* R.
> |
> | And in fact the promise was fulfilled.  I then started R and got:
> |
> | > R version 3.5.2 (2018-12-20) -- "Eggshell Igloo"
> | > Copyright (C) 2018 The R Foundation for Statistical Computing
> | > Platform: x86_64-pc-linux-gnu (64-bit)
> |
> | Ta-da!!!
> |
> | Was the problem essentially that I had been saying
> |
> | "sudo apt-get install r-base" rather than
> |
> | "sudo apt-get install r-base-core"?
> 
> No. One is a superset. A meta package.

Well, it worked.  And you can't argue with success.  It seems to have 
got the install procedure to put in all of the bits and pieces so that a 
working version of R was actually created.  E.g. the directory /etc/R 
(which I'd previously removed, hidden away) was created.  As
were /usr/bin/R, /usr/local/bin/R, the man files, etc.  (Which I had 
likewise previously hidden away.)

This did *not* happen before when I did "sudo apt-get install r-base".
So there was a significant and important impact from installing what you 
refer to as the "meta package".

> 
> These are all __Debian__ or __Ubuntu__ questions.  Be patient, learn some
> about your package manager.

I don't think so.  I have work to do; I don't wish to spend my time 
scrabbling around trying to find reliable and comprehensible 
documentation.  There is a huge amount of material on the web; much of 
it is out of date, inaccurate or misleading.   Generally what is in fact 
accurate is obscure and arcane in the extreme. If you know the answer 
already, you can find it on the web.  Otherwise not.

I expect a command to work, given that the syntax used is correct.

In this case I followed *EXACTLY* (and if you sense a tone of 
exasperation here, you sense correctly) the instructions given at

    https://cran.r-project.org/bin/linux/ubuntu/README.html

and they *did not work*.  Saying "Oh, you must have messed up your 
system" is not helpful.

> And this is not the correct list for Debian or Ubuntu basics so it might be
> polite of you to seek basic help elsewhere.

I did.  (Ubuntu Mate Community site, as I previously mentioned.  Then 
later "AskUbuntu".)  To no avail.

Although I think it is disingenuous to describe this as "basic" help. 
The problem seems to have been obscure.  If it was basic, why was no-one 
able to provide me with an answer?  It was really by pure serendipity 
(your use of dpkg -l rather than dpkg -L and my miss-interpretation of 
the output!!!) that I finally stumbled onto the solution.


<SNIP>

cheers,

Rolf

-- 
Honorary Research Fellow
Department of Statistics
University of Auckland
Phone: +64-9-373-7599 ext. 88276



More information about the R-SIG-Debian mailing list