[R-sig-Fedora] R 2.8.1

[Martyn Plummer]

On Sun, 2009-01-04 at 15:51 +0100, Peter Dalgaard wrote:
> Tom "spot" Callaway wrote:
> > On Sun, 2009-01-04 at 12:33 +0100, Peter Dalgaard wrote:
> > 
> >> Finally got around to testing it (sorry, but it didn't reach testing 
> >> till the 24th, and we have this thing call Christmas around here...).
> >>
> >> However, it now seems that I cannot update the karma system. Yesterday, 
> >> I got an "internal error" or some such when submitting, and now I can 
> >> submit, but only the text comes through, not my email addr. nor the 
> >> "works for me". So there's now two "anonymous tester" entries with a 
> >> bland smiley icon.
> > 
> > Hmm, they came through to my inbox with your name/email. Very odd. :)
> > 
> > ~spot
> > 
> 
> I spoke slightly too soon. This version has the same "missing mu" 
> syndrome as 2.8.0 did (and which occurs on SUSE too), try
> 
>   plot(1, type="n");text(1,1,quote(kappa+lambda+mu+nu+rho+sigma))
> 
> It doesn't happen when building on my own machine, so I can only assume 
> that something is up with the build machine or the chrooted build 
> environment.

I think we need to add pango-devel as a build requirement, so that the
cairo graphics device uses the pango layout engine.

If you want to test this, I have rebuilt the RPMS with mock, after
adding pango-devel to BuildRequires, and put them here:
http://calvin.iarc.fr/~martyn/R-2.8.1/

I don't get the same "missing mu" bug, but I have some commercial fonts
installed. I do get error messages about missing font metrics, however,
which would be due to the lack of pango.


I also noticed that libRmath-devel does not pull in the right release of
libRmath. The requirements for libRmath-devel are currently:

Requires: libRmath = %{version}, pkgconfig

but should be:

Requires: libRmath = %{version}-%{release}, pkgconfig


Martyn




-----------------------------------------------------------------------
This message and its attachments are strictly confidenti...{{dropped:8}}