On Thu, Dec 6, 2012 at 9:34 AM, Dirk Eddelbuettel <e...@debian.org> wrote:
| > [Are more Solaris-esque build slaves needed? Someone give a shout
> if so...
> | > we can sponsor some infrastructure there.]
> |
> | as long we (R community at large "in principle", de facto,
> | AFAIK, Prof Brian Ripley)
> | can have a Solaris testbed for CRAN which is reflected in the
> | CRAN package check, it would be really great if someone provided
> | a "Solaris-builder" server similar to the
> | win-builder.r-project.org one that Uwe (and his Depaertment/
> | University) provides.
>
> In a world with finite resources in terms of coding time and attention
> span,
> should we not provide such resources first for OS X and Linux? But maybe
> we
> need some empirics first so that we can rank-order compilers by relevance.
>
I would fully expect it to be the case that the distribution of build
issues would somewhat match the popularity of platforms. Modulated by
something like the approximate expertise on the platforms. E.g. you may
not see many problems from Solaris folks for a variety of reasons,
popularity being only one aspect of this... :-)
> I for one get more error emails from Intel icc users (say, two a year) than
> from Solaris users (probably about one or two, over all these years; and
> yes
> I excluded the CRAN maintainers here).
>
> That said, we do have a concrete offer for help from Elijah which should a)
> be commended ("nice job!") and used
I'm going to get 'official permission' from someone, in hopes that whatever
we do lasts longer. I'm expecting a positive reception, of course, or I
wouldn't have said it could be done. :-)
(oh - yes, I got uniquivocal permission to go full speed ahead. :-) )
Once I get that in hand, is B. Ripley the right person to reach out to?
I'm pretty unfamiliar with what the CRAN autobuilder world is like, these
days.
> but b) we should maybe try to think about
> what other resources would help.
>
Yup, this seems like a good time to bring that up, for sure. :-)
best,
--e
[[alternative HTML version deleted]]
______________________________________________
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
