spkg-check fails for the R spkg

[williamstein]

I build sage-4.0.2.alpha3 on sage.math with

export SAGE_CHECK="yes"

and it failed as follows:

...
Successfully installed r-2.6.1.p22
Running the test suite.
make[2]: Entering directory `/scratch/wstein/build/sage-4.0.2.alpha3/spkg/build/r-2.6.1.p22/src'
make[3]: Entering directory `/scratch/wstein/build/sage-4.0.2.alpha3/spkg/build/r-2.6.1.p22/src/tests'
make[4]: Entering directory `/scratch/wstein/build/sage-4.0.2.alpha3/spkg/build/r-2.6.1.p22/src/tests'
make[5]: Entering directory `/scratch/wstein/build/sage-4.0.2.alpha3/spkg/build/r-2.6.1.p22/src/tests/Examples'
make[6]: Entering directory `/scratch/wstein/build/sage-4.0.2.alpha3/spkg/build/r-2.6.1.p22/src/tests/Examples'
make[6]: Leaving directory `/scratch/wstein/build/sage-4.0.2.alpha3/spkg/build/r-2.6.1.p22/src/tests/Examples'
make[6]: Entering directory `/scratch/wstein/build/sage-4.0.2.alpha3/spkg/build/r-2.6.1.p22/src/tests/Examples'
collecting examples for package 'base' ...
make[7]: Entering directory `/scratch/wstein/build/sage-4.0.2.alpha3/spkg/build/r-2.6.1.p22/src/src/library'
 >>> Building/Updating help pages for package 'base'
     Formats: text html latex example
make[7]: Leaving directory `/scratch/wstein/build/sage-4.0.2.alpha3/spkg/build/r-2.6.1.p22/src/src/library'
make[6]: *** No rule to make target `0', needed by `base-Ex.Rout'.  Stop.
make[6]: Leaving directory `/scratch/wstein/build/sage-4.0.2.alpha3/spkg/build/r-2.6.1.p22/src/tests/Examples'
make[5]: *** [test-Examples-Base] Error 2
make[5]: Leaving directory `/scratch/wstein/build/sage-4.0.2.alpha3/spkg/build/r-2.6.1.p22/src/tests/Examples'
make[4]: *** [test-Examples] Error 2
make[4]: Leaving directory `/scratch/wstein/build/sage-4.0.2.alpha3/spkg/build/r-2.6.1.p22/src/tests'
make[3]: *** [test-all-basics] Error 1
make[3]: Leaving directory `/scratch/wstein/build/sage-4.0.2.alpha3/spkg/build/r-2.6.1.p22/src/tests'
make[2]: *** [check] Error 2
make[2]: Leaving directory `/scratch/wstein/build/sage-4.0.2.alpha3/spkg/build/r-2.6.1.p22/src'
Error while running the R testsuite ... exiting
*************************************Error testing package ** r-2.6.1.p22 **
*************************************
sage: An error occurred while testing r-2.6.1.p22
Please email sage-devel http://groups.google.com/group/sage-devel
explaining the problem and send the relevant part of

CC: @kcrisman

Component: build

Issue created by migration from https://trac.sagemath.org/ticket/6279

[dandrake]

comment:1

This fails for 4.1 on sage.math. Looking in the tests directory, I see a couple "Rout.fail" files, and both complain

Error in .find.package(package, lib.loc, verbose = verbose) : 
  there is no package called 'MASS'

Googling around, it looks like we'd need to install the recommended packages: https://stat.ethz.ch/pipermail/r-help/2002-May/021259.html. It looks like there is a "--without-recommended" flag for ./configure, but it may not do what we want: https://stat.ethz.ch/pipermail/r-devel/2008-April/049054.html

Sadly, you can't even get through the testsuite without the recommended packages.

[wjp]

comment:2

In the r-2.9.2 spkg, the failing tests are

src/library/utils/R-ex/capture.output.R
src/library/utils/R-ex/ls_str.R
src/library/grDevices/R-ex/densCols.R
src/library/graphics/R-ex/smoothscatter.R
src/library/stats/R-ex/add1.R
src/library/stats/R-ex/alias.R
src/library/stats/R-ex/anova.glm.R
src/library/stats/R-ex/aov.R
src/library/stats/R-ex/dummy.coef.R
src/library/stats/R-ex/extractAIC.R
src/library/stats/R-ex/glm.R
src/library/stats/R-ex/nls.R
src/library/stats/R-ex/stat.anova.R 
src/library/stats/R-ex/summary.glm.R

as well as the test for (R) issue #2586 in reg-tests-1.R and
the test labelled ## as.matrix.data.frame with coercion in reg-tests-3.R.

They seem to be caused by missing packages MASS, KernSmooth, survival.

Maybe we can just remove the failing tests?

[kcrisman]

comment:4

See #6235 for follow-up with a new spkg which apparently has different failures.

[kcrisman]

comment:5

Replying to @kcrisman:

See #6235 for follow-up with a new spkg which apparently has different failures.

I mean #6532. Also note this is only on some hardware.

[kcrisman]

comment:6

This may be fixed by #7485; see #6532 for details.

[kcrisman]

comment:7

I have tried this on a Linux and Macintel, and it seems that this is now fixed when one puts in all recommended packages and after #7485. I don't feel comfortable closing it, though, since I made the new spkg at #6532. Putting as "needs review", though it doesn't exactly...

[sagetrac-mvngu]

comment:8

Close as fixed by #6532.