r's install_packages is broken in many variants of sage

[williamstein]

See, e.g., this from sage-devel

I had a similar failure today, trying to:

r.install_packages("adapt")

after some fussing, runing ./sage as root, and using the notebook
interface I could get through the download phase, but same sorts of
failures in just as the gcc kicks in.  Seems several of the key R
scripts have "/home/wstein/..."  hard wired in to R_HOME_XXX, which
obviously will fail.  I tried editing the R startup scripts (among
others) but couldn't get it to work.

BTW, I installed from the latest Debian tarball into a Debian/VMWARE
machine just today.  So installing R packages is still an issue.

Since most of the value of R is the huge list of third party packages, it's very important that this get fixed.

CC: @jasongrout @sagetrac-mvngu

Component: packages: standard

Issue created by migration from https://trac.sagemath.org/ticket/4959

[kcrisman]

comment:1

See also #5246, which reports an extremely similar problem.

Incidentally, I have not had this problem at all in downloading a number of packages, and others have reported success lately, so it is possible this was specific to a certain release because of weird hard coding of links environment variables. This should be tested on a few different machines with packages that do not rely on the 'recommended' packages (see #2198 and #6532 for why).

[kcrisman]

comment:2

Incidentally, the R package above has been removed from R, it seems. It also seems this person was using the Debian install, which is now of course woefully out of date.

[kcrisman]

comment:3

On sage.math, after picking the mirror, I get one error message for trying to install 'abind' from the r console:

Warning message:
In doTryCatch(return(expr), name, parentenv, handler) :
  unable to load shared library '/home/.../sage-4.3.r0-sage.math.washington.edu-x86_64-Linux/local/lib/R//modules//R_X11.so':
  /scratch/mhansen/release/4.3/rc0/sage-4.3.rc0/local/lib/gcc-lib/x86_64-unknown-linux-gnu/4.0.3/libgcc_s.so.1: version `GCC_4.2.0' not found (required by /usr/lib/libstdc++.so.6)
Segmentation fault

but a different one when trying to use the r.install_packages() interface, which I unfortunately deleted.

[kcrisman]

comment:4

Here is the error message from that:

Warning messages:
1: In file.create(f.tg) :
  cannot create file '/scratch/mhansen/release/4.3/rc0/sage-4.3.rc0/local/lib/R/doc/html/packages.html', reason 'Permission denied'
2: In tools:::unix.packages.html(.Library) :
  cannot create HTML package index
> 

So again there seem to be problems with hard links in the R installation, since my binary wasn't in /scratch and certainly not in mhansen's scratch.

[kcrisman]

comment:5

And see here for reports on several other platforms.

[kcrisman]

comment:6

The behavior with the version `GCC_4.2.0' not found (required by /usr/lib/libstdc++.so.6) continues to be the case on boxen.math after the upgrade at #6532. On the plus side, it no longer refers to someone else's installation.

Here is another interesting error which might help us - is it possible that a number of R packages require non-universal libraries?
{{{* installing source package ‘rgl’ ...
checking for gcc... gcc -std=gnu99
checking for C compiler default output file name... a.out
checking whether the C compiler works... yes
checking whether we are cross compiling... no
checking for suffix of executables...
checking for suffix of object files... o
checking whether we are using the GNU C compiler... yes
checking whether gcc -std=gnu99 accepts -g... yes
checking for gcc -std=gnu99 option to accept ISO C89... none needed
checking how to run the C preprocessor... gcc -std=gnu99 -E
checking for gcc... (cached) gcc -std=gnu99
checking whether we are using the GNU C compiler... (cached) yes
checking whether gcc -std=gnu99 accepts -g... (cached) yes
checking for gcc -std=gnu99 option to accept ISO C89... (cached) none needed
checking for libpng-config... yes
configure: using libpng-config
configure: using libpng dynamic linkage
checking for X... libraries /usr/lib, headers /usr/include
checking GL/gl.h usability... no
checking GL/gl.h presence... no
checking for GL/gl.h... no
checking GL/glu.h usability... no
checking GL/glu.h presence... no
checking for GL/glu.h... no
configure: error: missing required header GL/gl.h
ERROR: configuration failed for package ‘rgl’
}}}
This was on boxen.math; on my own Mac I did not have any problems with this package (a package installed by the package 'depth', which I needed for some other computations). This tells me that

[kcrisman]

comment:7

With the Fortran issues in #7485 cleared up, #6532 seems to now work to fix this. We can definitely get rid of the warning for OSX! Look on #7521 for that.

[sagetrac-mvngu]

comment:8

Close as fixed by #6532.