fix build failure on R-Forge?

[jaganmn]

I realize that GitHub is the new place for glmmADMB development, but note that R-Forge currently supplies an "empty" version 0.8.3.3:

> install.packages("glmmADMB", repos = "http://R-Forge.R-project.org")
Installing package into ‘/Users/mikael/Desktop/Matrix-reverse/preCRAN/preCRAN/Library’
(as ‘lib’ is unspecified)
trying URL 'http://R-Forge.R-project.org/src/contrib/glmmADMB_0.8.3.3.tar.gz'
Content type 'application/x-gzip' length 999 bytes
==================================================
downloaded 999 bytes

* installing *source* package ‘glmmADMB’ ...
** using staged installation
** help
No man pages found in package  ‘glmmADMB’ 
*** installing help indices
** building package indices
** testing if installed package can be loaded from temporary location
** testing if installed package can be loaded from final location
** testing if installed package keeps a record of temporary installation path
* DONE (glmmADMB)

The downloaded source packages are in
    ‘/private/var/folders/n7/v9s56rmd5hn17d3f1qj13l7m0000gn/T/RtmpGS3fhc/downloaded_packages’
> names(asNamespace("glmmADMB"))
[1] ".packageName"         ".__NAMESPACE__."      ".__S3MethodsTable__."

It would be useful to have at least a working build there, even if it is not the most recent. I installed this useless version 0.8.3.3 without realizing, then spent a long time trying to understand why example("glmmadmb_tidiers", package = "broom.mixed") no longer worked ...

[bbolker]

This might be an R-forge glitch.

Doing

svn checkout svn+ssh://bbolker@scm.r-forge.r-project.org/svnroot/glmmadmb/ glmmadmb_svn
cd glmmadmb_svn/www/repos/src/contrib
Rscript -e "install.packages('glmmADMB_0.8.3.3.tar.gz'); length(names(asNamespace('glmmADMB')))"

seems to work fine.

I could have sworn I got it to work via install.packages(..., repos = ...) once, but now I can't repeat it.

Trying the following out of stubbornness:

f <- function(delay = 10) {
   Sys.sleep(delay)
  cc <- capture.output(type = "message", install.packages("glmmADMB", repos = "http://R-Forge.R-project.org"))
  grep("downloaded [0-9]+", value = TRUE, cc )
}
res <- replicate(50, f())

... ugh, it failed ("downloaded 999 bytes") every time ...

this page says "File length: 9391177 byte(s)" ...

[jaganmn]

The problem is this file: http://R-Forge.R-project.org/src/contrib/glmmADMB_0.8.3.3.tar.gz. You can trigger a rebuild by committing to the SVN repository. Maybe that will result in a reasonable tarball being copied into R-Forge.R-project.org/src/contrib. I'm not actually sure what happens behind the scenes.

[bbolker]

I tried a dummy commit, let's see what happens.

I don't think I realized that R-forge would automatically build packages for the repo (as noted in the doc); IIRC I had been building them myself and committing to the SVN.

I could have sworn that I did succeed once with installing directly from the repo earlier today, but of course now I can't replicate that ...

[bbolker]

Nothing obvious in the Linux build log, except that it's ancient:

Fri Feb 23 19:05:17 2018: Building tarball for package glmmADMB (SVN revision 289)
using R version 3.4.3 Patched (2018-02-21 r74284) ...

* checking for file ‘glmmADMB/DESCRIPTION’ ... OK
* preparing ‘glmmADMB’:
* checking DESCRIPTION meta-information ... OK
* compacting vignettes and other PDF files
* checking for LF line-endings in source and make files and shell scripts
* checking for empty or unneeded directories
Removed empty directory ‘glmmADMB/tests’
* looking to see if a ‘data/datalist’ file should be added
* re-saving image files
* building ‘glmmADMB_0.8.3.4.tar.gz’

Run time: 4.2 seconds.

The Windows build failed because R2admb was unavailable.

Hopefully the SVN commit will clear all this up; don't look forward to going down the rabbit hole ...

[jaganmn]

The current build/check status will be updated in the R Packages tab of the R-Forge project page. If it seems to stall or fail after waiting a day or two, then maybe just ask the R-Forge admin to do the rabbit hole part.

[jaganmn]

The build failed. I would just e-mail R-Forge@R-Project.org...