[R-sig-Debian] Ubuntu 64

[Chris Evans]

Johannes Ranke sent the following  at 11/08/2007 20:07:
> Hi Chris, hi list
> 
> * Chris Evans <chrishold at psyctc.org> [070802 09:00]:
>> Emmanuel Charpentier sent the following  at 26/07/2007 15:18:
>>> Dear sig-R-Debian
>>>
>>> I'd like to second Christophe Bonenfan's plea for an up-to-date Ubuntu
>>> 64 (and maybe an up-to-date Debian 64) repository(ies). There seems to
>> ... and for me it's particularly Debian 64 stable.  I'm more
>> psychotherapist than dental surgeon ... but otherwise Emmanuel's
>> self-description fits well!
> 
> Are there any problems with the amd64 backports for Debian stable on
> CRAN that I should know of?

Very belated reply to this.  First of all, many, many thanks for all
you're doing Johannes.

I've only just had time to come back to using R on my server and I
realised I was being a bit dumb and that the Debian upgrade policy and
the remarkable generosity of the R project probably meant I'd have to
tweak /etc/apt/sources.list and point at some source other than the main
Debian stable one.

Two points that may help others though:
1) The crucial information about how to do it is in CRAN, for my local
depository it's at:

http://www.stats.bris.ac.uk/R/bin/linux/debian/

I think the frames system used by CRAN can make the URL you're at a bit
unclear (Firefox shows I'm at http://www.stats.bris.ac.uk/R/ as the key
page is an frame on that page, one has just been jumped to a single page
of http://www.stats.bris.ac.uk/R/bin/linux/debian/

What it says there is that you need:
"deb http://<favorite-cran-mirror>/bin/linux/debian stable/"
in your /etc/apt/sources.list

I, perhaps stupidly, put in:
"deb http://www.stats.bris.ac.uk/bin/linux/debian stable/"

and got an file not found error as it's actually:
"deb http://www.stats.bris.ac.uk/R/bin/linux/debian stable/"

I think either the listing of the CRAN mirror or that line in that file
need to be changed to make mapping clearer.

2) the other problem I hit was that not all the packages seem to have
CRC values stored for them so you get a warning from dselect and have to
realise that the default is _not_ to ignore it, to be more precise, I
had to confirm twice in effect that was ignoring it.

However, up and running like a dream and interestingly it immediately
solved a problem I was having with sink(): clearly it had been a bug and
had been fixed!  Wonderful!

Many thanks Johannes and sorry if I spread confusion!

Very best all,

Chris


-- 
Chris Evans <chris at psyctc.org> Skype: chris-psyctc
Professor of Psychotherapy, Nottingham University;
Consultant Psychiatrist in Psychotherapy, Notts PDD network;
Research Programmes Director, Nottinghamshire NHS Trust;
*If I am writing from one of those roles, it will be clear. Otherwise*
*my views are my own and not representative of those institutions    *