[R-sig-ME] Failure to load lme4 on Mac

[Ken Knoblauch]

Simon Urbanek <simon.urbanek at ...> writes: 
> On Jul 15, 2010, at 7:54 PM, Douglas Bates wrote:
> 
> > On Thu, Jul 15, 2010 at 6:47 PM, Daniel Myall
> <daniel.lists at ...> wrote:
> >> I've attempted to reproduce the error seen on the CRAN on several 
Mac OS X
> >> machines and haven't been able to. I'm able to build the CRAN 
lme4 version
> >> and the svn versions of lme4 and lme4a without issues.
> > 
> >> There is obviously an interaction with lme4 and the version of 
xcode
> >> installed on the CRAN Mac OS build machine that is resulting in the 
test
> >> failure. Is there a page that details the setup of the Mac OS X CRAN 
build
> >> machine?
> > 
> >> I'm happy to determine what is going on and attempt to resolve this 
(it will
> >> mean I don't have to custom compile lme4 for users all the time).
> > 
> > Thank you Daniel.   Simon Urbanek is the person who coordinates the
> > Mac OS X builds so it might be best if you could share your
> > experiences on successful building with him and perhaps determine 
why
> > the CRAN builds are encountering the error.  I think it unlikely that
> > it would be due to an out-of-date Xcode compiler set as usually 
Simon
> > is one of the people working with the most recent tools.
> > 
> 
> The CRAN machine is Mac OS X 10.5.8 with Xcode 3.1.4 (using 
gcc-4.2) 
which is 
indeed the latest 
Xcode version
> for Leopard. If you have any hints as of why lme4 may be failing they 
are 
highly appreciated.
> 
> Thanks,
> Simon
> 
> > I believe that Simon, who I have cc'd on this response, may be on
> > vacation or otherwise away from his email this week.  I suggest that
> > you pursue the issue with him off-list when he is available again.
> > 
> >> 
> >> Daniel
> >> 
> >> 
> >> On 16/07/10 10:13 AM, Douglas Bates wrote:
> >>>>> 
> >>>>> Neither Martin nor I use Mac OS and we can't reproduce the 
problem on
> >>>>> other operating systems so we are kind of stuck.
> >>>>> 
> >>>>> Apparently it is a matter of one of the tests failing but without
> >>>>> access to a system that exhibits the failure I wouldn't know 
where to
> >>>>> start debugging.  Is some kind person using Mac OS were to 
download
> >>>>> and test the package and provide some details about the 
nature of the
> >>>>> failure we might be able to get started.
> >>>>> 
Just having updated Xcode to 3.1.4 and fixed the versions of gcc, etc. so
that I'm able to install latest version of Rcpp and lme4a without any 
problems,
I'll take a bite, but maybe I'm just repeating what you have already seen.
I can install lme4 from source from Rforge with no problem.
Just for starters here is my sessionInfo
R version 2.11.1 Patched (2010-07-15 r52539) 
i386-apple-darwin9.8.0 

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

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

other attached packages:
[1] lme4_0.999375-34   Matrix_0.999375-41 lattice_0.18-8    

loaded via a namespace (and not attached):
[1] grid_2.11.1   nlme_3.1-96   stats4_2.11.1

I dowloaded lme4_0.999375-34.tar.gz from CRAN and ran
R CMD check on it and came up with
* checking tests ...
  Running ‘extras.R’
  Running ‘lmer-1.R’
 ERROR
Running the tests in 'tests/lmer-1.R' failed.
Last 13 lines of output:
  +                  x3=rnorm(20,1))
  > m0 <- lmer(y ~ (x1 + x2)|ff, data = D)
  > m1 <- lmer(y ~ x1 + x2|ff  , data = D)
  > m2 <- lmer(y ~ x1 + (x2|ff), data = D)
  > m3 <- lmer(y ~ (x2|ff) + x1, data = D)
  > stopifnot(identical(ranef(m0), ranef(m1)),
  +     identical(ranef(m2), ranef(m3)),
  +    inherits(tryCatch(lmer(y ~ x2|ff + x1, data = D), error = function(e)e),
  +                    "error"))
  Error: identical(ranef(m0), ranef(m1)) is not TRUE
  In addition: Warning message:
  In Ops.factor(ff, x1) : + not meaningful for factors
  Execution halted
which I think is the reported problem.

I went into the file lmer-1.R which seems to be where this is
coming from and sourced the lines indicated above and get

> D <-  data.frame(y= rnorm(20,10), ff = gl(4,5),
+                  x1=rnorm(20,3), x2=rnorm(20,7),
+                  x3=rnorm(20,1))
> m0 <- lmer(y ~ (x1 + x2)|ff, data = D)
> m1 <- lmer(y ~ x1 + x2|ff  , data = D)
> m2 <- lmer(y ~ x1 + (x2|ff), data = D)
> m3 <- lmer(y ~ (x2|ff) + x1, data = D)
> stopifnot(identical(ranef(m0), ranef(m1)),
+           identical(ranef(m2), ranef(m3)),
+           inherits(tryCatch(lmer(y ~ x2|ff + x1, data = D), error = 
function(e)e),
+                    "error"))
Error: identical(ranef(m2), ranef(m3)) is not TRUE
In addition: Warning message:
In Ops.factor(ff, x1) : + not meaningful for factors
> identical(ranef(m0), ranef(m1))
[1] TRUE
> identical(ranef(m2), ranef(m3))
[1] FALSE

so it happens in an R session as well.
Checking the ranef's
> ranef(m2)
$ff
  (Intercept)          x2
1  -1.4142222  0.15618464
2   0.4787662 -0.05287424
3   0.3193410 -0.03526755
4   1.7224362 -0.19022335

> ranef(m3)
$ff
  (Intercept)          x2
1  -1.4141930  0.15618142
2   0.4787944 -0.05287736
3   0.3193632 -0.03527000
4   1.7224652 -0.19022656

They do indeed differ.  I'm on a 32 bit machine, could this be
a 32/64 bit difference or something at the level of BLAS/ATLAS level?
That's out of my league.  If I have time, I'll try it on a 64 bit Mac, but I'm
not likely to get back to that before about the beginning of August.


Hope that this is useful.

Ken

-- 
Ken Knoblauch
Inserm U846
Stem-cell and Brain Research Institute
Department of Integrative Neurosciences
18 avenue du Doyen Lépine
69500 Bron
France
tel: +33 (0)4 72 91 34 77
fax: +33 (0)4 72 91 34 61
portable: +33 (0)6 84 10 64 10
http://www.sbri.fr/members/kenneth-knoblauch.html