[Rd] missing space in R version specifier makes PACKAGES file unreadable by install.packages()

Duncan Murdoch murdoch.duncan at gmail.com
Thu Mar 21 18:28:55 CET 2013


On 21/03/2013 12:59 PM, Dan Tenenbaum wrote:
> On Thu, Mar 21, 2013 at 9:52 AM, Dan Tenenbaum <dtenenba at fhcrc.org> wrote:
> > Hi,
> >
> > After updating to R-3.0 beta r62328, I get the following:
> >
> >> install.packages("Biobase", type="source", repos="http://george2/BBS/2.12/bioc")
> > Error in do.call(op, list(v_c, v_t[[op]])) :
> >   could not find function "R (>=2.15.1)"
> >
> > The problem can be fixed by adding a space after >= in the offending
> > package's DESCRIPTION file and re-generating the PACKAGES file with
> > tools:::write_PACKAGES().
> >
> > However, this worked OK in r62077.
>
> Beg pardon, this did not work in r62077 either.
>
> However I still think it's worth reporting.

I don't think this has changed recently.  The manual currently says "The 
comment should contain a comparison operator, whitespace and a valid 
version number."  In 2.15.0 it was less explicit, but the example 
contained a space.  It may be that some code path worked before more or 
less by accident, because in December changes were made to the parsing 
to avoid a tricky reentrancy problem, but I don't think it has ever been 
documented to work without whitespace.

I can't think of the reason white space was required, but the code does 
explicitly look for it, so I'd be reluctant to change it.

Duncan Murdoch

> Dan
>
>
> > I'm not sure if >=2.15.1 (without
> > the space) is valid syntax, but I wonder if R can be a bit more
> > forgiving, as this issue derailed our daily builds.
> >
> >> sessionInfo()
> > R version 3.0.0 alpha (2013-03-18 r62312)
> > Platform: x86_64-apple-darwin10.8.0 (64-bit)
> >
> > locale:
> > [1] C
> >
> > attached base packages:
> > [1] stats     graphics  grDevices utils     datasets  methods   base
> >
> > loaded via a namespace (and not attached):
> > [1] tools_3.0.0
> >
> >
> > Thanks,
> > Dan
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel



More information about the R-devel mailing list