[ESS-bugs] ESS 13.09-1 with old R versions fails pretty badly

[Rodney Sparapani]

On 10/29/2013 03:26 PM, Vitalie Spinu wrote:
> It is more complex than this. R code is everywhere nowadays. Let's not
> go in that domain, it will be a pain to maintain and explain to the
> users. We already have install related questions more often than we
> want.
>
> I appreciate that Martin started working on that, as I don't have the
> necessary tools and experience to deal with such old R versions, but
> please keep in mind MELPA users and people (like me) who don't use
> make.
>
> That essentially means that you have to ship ESSR_<version>.tar.gz with
> ESS if you want to go that route. Keeping installed copy of ESSR in SVN
> is better IMO, and it looks like Martin is going in that direction.
>
>     Vitalie
>
> PS: Also, don't forget folks on remotes:) Then need to be able to
> install from a web source.

Well, I am not forgetting.  But, all of the various ways of installing
ESS are creating havoc (not me ;o)  Many times the Makefile is out of
sync.  So, without some sort of automated testing, is there anything
else we can do to unite these different install methods?  I understand
why we are getting questions because I don't like it either!  I realize
that R packages have problems like this too (and emacs packages).  But,
I would love to hear a suggestion on how we can make this more uniform.
Just to be clear we have source, MELPA, github, SVN, Vincent's distro
and various flavors of Linux distros (Fedora, Debian, etc.) and God
knows what.  This is too much for our little project to realistically
manage.  I don't want to support all of that.  It's hard enough
fixing ESS bugs on my own system let alone trying to guess what somebody
else has done.

-- 
Rodney Sparapani, PhD
Manager of Statistical & Computational Operations
Center for Patient Care and Outcomes Research (PCOR)
Medical College of Wisconsin (MCW), Milwaukee, USA
http://www.linkedin.com/in/rodneysparapani