On Fri, 23 Jan 2009, Dirk Eddelbuettel wrote:


reassign 510251 postgresql-8.3-plr
thanks

Andreas,

First off, thanks for maintaining pl-r. I wasn't aware that it has landed in
your lap. I was aware that the Pg maintainers didn't really want it (for lack
of R expertise) but I had no spare capacity to take it on myself.

On 23 January 2009 at 14:10, Andreas Tille wrote:
| On Fri, 23 Jan 2009, Andreas Tille wrote:
|
| > You can work around this problem for your local installation by just
| > copying src/include/Rdevices.h into the plr source directory.  The
| > you can build the package for the moment.
|
| I have to revert this.

Yes, that was plain wrong.  R (upstream) changed their interface and
tightened this. It also affected things I maintain such as the RPy package.

So the error is plainly with the client package. It is not an R issue
inasmuch as that these interface were never public (yet pl/R is a well-known
example even used in the 'R Extensions' manual about embedding).

| I learned that Rgraphics.h is lacking as well
| and once you include these both in the plr source tree and try to
| build against R 2.8 you get something like:

Did you talk Joe Conway about this?  Joe as upstream author has always been
very responsive clueful when I contacted him about pl/R (which I had played
with over the years).

Now the final question for Faheem: Pl/R has existed in Debian for years. Why
did you think you needed to rebuild it for etch?  Were you lacking a
particular combination of versions of R and/or Pg on etch ?

Hi Dirk, Andreas, Joe,

Thanks for picking up the ball on this one. Dirk, in answer to your specific question, I needed to have pl/r installed on a particular machine which I do not administer. This machine is running etch, and also your backport of R 2.8.

ii r-base 2.8.1-1~etchcran.0 GNU R statistical computation and graphics s ii r-base-core 2.8.1-1~etchcran.0 GNU R core of statistical computation and gr

So, you see why I needed to build pl/r against etch and that particular version of R.

I'm sure we can sort all this out.

I hope so too. Writing to the pl/r mailing list was on my todo list (in fact I would probably have got to it in the next few days if I had not heard from Andreas), but since Joe is now in the loop, it is no longer necessary. I'm just surprised this was not a known issue already.

                                                        Regards, Faheem.



--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Reply via email to