The R-multcomp maintainer asked me to look into why it was failing to
build in rawhide. When I looked into it, I discovered that something is
broken in how R CMD CHECK --no-install works in 2.10.1. I couldn't tell
you what is broken exactly, but I took the time today to package up all
the Suggests (and their depchain) and tested the package without
--no-install (and without --no-latex) and it works fine again.
I asked Orion (the R-multcomp maintainer) if he was interested in
maintaining/co-maintaining/reviewing these new R packages, but he wasn't
interested. In fact, he said he was more inclined to orphan R-multcomp.
Is there any interest in helping me maintain and review these new R
packages in Fedora? The packages are already done, its just the reviews
and upkeep that I'd need help with.
Here's the packages:
These are the R packages I made today:
coin, colorspace, ipred, lme4, mboost, mlbench, modeltools, multicore,
party, robustbase, sandwich, strucchange, vcd, xtable
If any of these look interesting to you, speak up! If no one cares, I
might just let R-multcomp orphan off, but if there are some willing
helpers, I'll keep it alive with its new dependencies.
I have been quickly playing with  to generate a kind of report on the
package that would need work.
There is the report:
spot -- R-nws is outdated, 126.96.36.199 vs 188.8.131.52 upstream
denisarnaud -- R-msm is outdated, 0.9.5 vs 0.9.6 upstream
orion -- R-car is outdated, 1.2 vs 1.2.16 upstream
orion -- R-lmtest is outdated, 0.9 vs 0.9.26 upstream
orion -- R-multcomp is outdated, 1.1 vs 1.1.6 upstream
orion -- R-mvtnorm is outdated, 0.9 vs 0.9.9 upstream
orion -- R-systemfit is outdated, 1.1 vs 1.1.4 upstream
orion -- R-zoo is outdated, 1.6 vs 1.6.2 upstream
pingou -- R-Biostrings is outdated, 2.14.10 vs 2.14.12 upstream
jamatos -- R-mAr is outdated, 1.1.1 vs 1.1.2 upstream
jamatos -- R-waveslim is outdated, 1.6.1 vs 1.6.3 upstream
nigelj -- R-RSQLite is outdated, 0.8.1 vs 0.8.3 upstream
Total: 12 packages outdated/needing some work
I believe all the packages of orion are false positive (it is the issue
of the - in the version number).
I believe R-nws is also wrong, but the page on the cran is not been
updated since 184.108.40.206.
We could imagine a monthly reminder if you are interested.