[Rd] Unexpected behaviour with download.packages on Windows

[Duncan Murdoch]

On 29/09/2017 12:44 PM, Rui Barradas wrote:
> Hello,
> 
> The help pages for download.packages and install.packages say
> 'type = getOption("pkgType")'.
> And on Windows I get
> 
> getOption("pkgType")
> [1] "both"
> 
> which means in ?getOptions
> 
> pkgType:
> 
>       The default type of packages to be downloaded and installed – see
> install.packages. Possible values are "win.binary", "source" and "both"
> (the default).
> 
> Are you sure the package you're downloading has a binary version or is
> it source only?

The claim is right -- download.packages() interprets "both" as "source". 
  The documentation is unclear about this, but the source is pretty 
simple.  The source to install.packages() is a lot more complicated, but 
I believe it would normally interpret "both" as "win.binary" on Windows, 
as documented.

Duncan Murdoch

> 
> sessionInfo()
> R version 3.4.1 (2017-06-30)
> Platform: x86_64-w64-mingw32/x64 (64-bit)
> Running under: Windows 7 x64 (build 7601) Service Pack 1
> 
> Matrix products: default
> 
> locale:
> [1] LC_COLLATE=Portuguese_Portugal.1252
> LC_CTYPE=Portuguese_Portugal.1252
> [3] LC_MONETARY=Portuguese_Portugal.1252 LC_NUMERIC=C
> 
> [5] LC_TIME=Portuguese_Portugal.1252
> 
> attached base packages:
> [1] stats     graphics  grDevices utils     datasets  methods   base
> 
> other attached packages:
>    [1] Hmisc_4.0-3          Formula_1.2-2        survival_2.41-3
>    [4] lattice_0.20-35      DescTools_0.99.22    RecordLinkage_0.4-10
>    [7] ffbase_0.12.3        ff_2.2-13            bit_1.1-12
> [10] RSQLite_2.0          DBI_0.7              tcR_2.2.1.11
> [13] igraph_1.1.2         reshape2_1.4.2       gridExtra_2.2.1
> [16] dplyr_0.7.1          ggplot2_2.2.1        sos_2.0-0
> [19] brew_1.0-6
> 
> loaded via a namespace (and not attached):
>    [1] Rcpp_0.12.11        stringdist_0.9.4.6  mvtnorm_1.0-6
>    [4] class_7.3-14        assertthat_0.2.0    digest_0.6.12
>    [7] ipred_0.9-6         R6_2.2.2            plyr_1.8.4
> [10] backports_1.1.0     acepack_1.4.1       ada_2.0-5
> [13] e1071_1.6-8         rlang_0.1.1         lazyeval_0.2.0
> [16] data.table_1.10.4   blob_1.1.0          rpart_4.1-11
> [19] Matrix_1.2-10       checkmate_1.8.3     splines_3.4.1
> [22] stringr_1.2.0       foreign_0.8-69      htmlwidgets_0.9
> [25] munsell_0.4.3       compiler_3.4.1      pkgconfig_2.0.1
> [28] base64enc_0.1-3     manipulate_1.0.1    htmltools_0.3.6
> [31] nnet_7.3-12         evd_2.3-2           htmlTable_1.9
> [34] tibble_1.3.3        prodlim_1.6.1       expm_0.999-2
> [37] MASS_7.3-47         grid_3.4.1          xtable_1.8-2
> [40] gtable_0.2.0        magrittr_1.5        scales_0.4.1
> [43] stringi_1.1.5       bindrcpp_0.2        latticeExtra_0.6-28
> [46] boot_1.3-19         fastmatch_1.1-0     lava_1.5.1
> [49] RColorBrewer_1.1-2  tools_3.4.1         bit64_0.9-7
> [52] glue_1.1.1          parallel_3.4.1      colorspace_1.3-2
> [55] cluster_2.0.6       memoise_1.1.0       knitr_1.16
> [58] bindr_0.1
> 
> 
> Hope this helps,
> 
> Rui Barradas
> 
> Em 29-09-2017 16:00, Hong Ooi via R-devel escreveu:
>> If no 'type' is specified, download.packages("pkgname") will download source packages (.tar.gz files), even on Windows. However, the help says
>>
>>
>>     type   character string, indicate which type of packages: see install.packages.
>>
>>
>> and on Windows, install.packages defaults to downloading binary packages.
>>
>> Is this intended behaviour on the part of download.packages? This is on R 3.3.3 and 3.4.1; I haven't tested on 3.4.2 but there's no indication the function has changed.
>>
>> ______________________________________________
>> R-devel at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-devel
>>
> 
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>