[R-SIG-Mac] lme4/macos crossover question

[David Winsemius]

On Aug 14, 2015, at 9:52 AM, Ben Bolker wrote:

> 
> As pointed out in this lme4 issue: https://github.com/lme4/lme4/issues/302 ,
> lme4 is behaving badly on Snow Leopard, failing to find an
> (apparently) exported function
> 
> urlhead <- "https://www.r-project.org/nosvn/R.check/r-"
> urltail <- "/gamm4-00check.html"
> versions <- c("devel-osx-x86_64-clang",
>              "release-osx-x86_64-mavericks",
>              "release-osx-x86_64-snowleopard")
> checkfun <- function(ver) {
>    uu <- paste0(urlhead,ver,urltail)
>    rr <- readLines(url(uu,method="libcurl"))
>    any(grepl("could not find function.*REMLcrit",rr))
> }
> sapply(versions,checkfun)

I could not get those to run with method="libcurl", possibly since I have not yet updated to 3.2.2? This did run:

 checkfun <- function(ver) {
    uu <- paste0(urlhead,ver, urltail)
    rr <- readLines(url(uu))
    any(grepl("could not find function.*REMLcrit",rr))
 }

 sapply(versions,checkfun)

        devel-osx-x86_64-clang   release-osx-x86_64-mavericks 
                         FALSE                          FALSE 
release-osx-x86_64-snowleopard 
                          TRUE 

> 
> Here's the (broken to make Gmane happy) link to the bad case:
> 
> https://www.r-project.org/nosvn/R.check/
>      r-release-osx-x86_64-snowleopard/gamm4-00check.html
> 
> The REMLcrit function is exported, at least in all recent versions of
> the package (and we haven't been able to reproduce on any other
> platforms).
> 
> Any ideas ... ?  Possibly the Snow Leopard test build on CRAN is
> using an out-of-date version of lme4 ... ?
> 

R version 3.2.1 (2015-06-18) -- "World-Famous Astronaut"
Platform: x86_64-apple-darwin10.8.0 (64-bit)

>   sessionInfo()
R version 3.2.1 (2015-06-18)
Platform: x86_64-apple-darwin10.8.0 (64-bit)
Running under: OS X 10.7.5 (Lion)

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] grDevices utils     datasets  stats     graphics  grid      methods   base     

other attached packages:
 [1] lme4_1.1-8        Matrix_1.2-2      stringr_1.0.0     dplyr_0.4.2      
 [5] ada_2.0-3         rpart_4.1-10      MASS_7.3-43       digest_0.6.8     
 [9] segmented_0.5-1.1 rms_4.3-1         SparseM_1.6       Hmisc_3.16-0     
[13] ggplot2_1.0.1     Formula_1.2-1     survival_2.38-3   sos_1.3-8        
[17] brew_1.0-6        lattice_0.20-33  

loaded via a namespace (and not attached):
 [1] Rcpp_0.12.0         nloptr_1.0.4        RColorBrewer_1.1-2  plyr_1.8.3         
 [5] tools_3.2.1         polspline_1.1.12    gtable_0.1.2        nlme_3.1-121       
 [9] DBI_0.3.1           parallel_3.2.1      mvtnorm_1.0-3       proto_0.3-10       
[13] gridExtra_2.0.0     cluster_2.0.3       nnet_7.3-10         R6_2.1.0           
[17] foreign_0.8-65      multcomp_1.4-1      minqa_1.2.4         latticeExtra_0.6-26
[21] TH.data_1.0-6       reshape2_1.4.1      magrittr_1.5        scales_0.2.5       
[25] codetools_0.2-14    splines_3.2.1       assertthat_0.1      colorspace_1.2-6   
[29] labeling_0.3        quantreg_5.11       sandwich_2.3-3      stringi_0.5-5      
[33] acepack_1.3-3.3     lazyeval_0.1.10     munsell_0.4.2       zoo_1.7-12         

[R.app GUI 1.66 (6956) x86_64-apple-darwin10.8.0]

I installed today an updated macbinary version of lme4, lme4_1.1-8.tgz, which the GUI package installer did for me with all the many dependencies. That is the latest version at CRAN for SL:

http://cran.cnr.berkeley.edu/bin/macosx/contrib/3.2/

Is that version "out of date"?  I suppose it might be, since I see this at the top of the CRAN package check:

• using R version 3.2.0 Patched (2015-04-28 r68272)
• using platform: x86_64-apple-darwin10.8.0 (64-bit)
• using session charset: UTF-8
• checking for file ‘gamm4/DESCRIPTION’ ... OK
• this is package ‘gamm4’ version ‘0.2-3’

I get this at the console:

>  require(lme4) 
Loading required package: lme4
Loading required package: Matrix
> set.seed(0) 
> y <- rnorm(20)
> x <- rnorm(20) 
> g <-factor(rep(1:2,10)) 
> REMLcrit(lmer(y~(1|g)+x)) 
[1] 57.98661

I'd be happy to remove and attempt reinstalling from source something else if that would be of any use to you wiser folk, who know what you are looking for and only need cannon fodder, er, willing test subjects.


>  thanks
>    Ben Bolker
> 
> _______________________________________________
> R-SIG-Mac mailing list
> R-SIG-Mac at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-mac

David Winsemius
Alameda, CA, USA