"# install.packages" ?

[LorenAmelang]

Installation
You can install the development version from [GitHub](https://github.com/) with:
# install.packages("devtools")

Yes, I'm ignorant about R, and I've jumped into the thick of it as I often do. To me, the '#' before that command means it needs to be entered at the Linux prompt with sudo... Does it mean something else to R fans? I don't see it in other instructions...

Quickly learned that didn't work. But I got confusing messages entering the command in R as well. Turns out devtools is not installed with R, and after hours of trying to install it showed:

ERROR: dependencies ‘usethis’, ‘httr’, ‘rcmdcheck’, ‘roxygen2’, ‘rversions’ are not available for package ‘devtools’
* removing ‘/home/pi/R/arm-unknown-linux-gnueabihf-library/4.0/devtools’
There were 12 warnings (use warnings() to see them)

I then found the remotes package, which installed relatively quickly and downloaded webmorphR, but it choked compiling Rcpp. Maybe we just can't do this on a Raspberry Pi?

(I love learning new things...)

Hmmm... Still building webmorphR, and getting lots of reports like:

/usr/include/c++/10/bits/vector.tcc:426:7: note: parameter passing for argument of type ‘std::vector<Catch::SectionEndInfo>::iterator’ changed in GCC 7.1
  426 |       vector<_Tp, _Alloc>::
      |       ^~~~~~~~~~~~~~~~~~~
In file included from /usr/include/c++/10/vector:67,
                 from testthat/vendor/catch.h:671,
                 from testthat/testthat.h:65,
                 from test-runner.cpp:8:

Do those mean this isn't going to work even if it does eventually finish? Does it require some specific version of GCC that isn't automatically selected?

[debruine]

The # is the R comment character. It’s just a note on the code to to install devtools (or remotes) in R if you run the install code and find that you don’t have devtools installed yet. You usually need command like tools like make to install devtools. I googled it and it looks like you can install R on a raspberry pi, but I don’t know anything about it. It does look like there might be some problems with Rccp, which is required for interacting with magick. You’ll also have to install magick manually if you aren’t using Mac or Windows (I haven’t included instructions because Linux users are good at trouble-shooting). See https://docs.ropensci.org/magick/articles/intro.html#build-from-source

…

On 22 May 2022, at 06:33, LorenAmelang ***@***.***> wrote:  Installation You can install the development version from [GitHub](https://github.com/) with: # install.packages("devtools") Yes, I'm ignorant about R, and I've jumped into the thick of it as I often do. To me, the '#' before that command means it needs to be entered at the Linux prompt with sudo... Does it mean something else to R fans? I don't see it in other instructions... Quickly learned that didn't work. But I got confusing messages entering the command in R as well. Turns out devtools is not installed with R, and after hours of trying to install it showed: ERROR: dependencies ‘usethis’, ‘httr’, ‘rcmdcheck’, ‘roxygen2’, ‘rversions’ are not available for package ‘devtools’ * removing ‘/home/pi/R/arm-unknown-linux-gnueabihf-library/4.0/devtools’ There were 12 warnings (use warnings() to see them) I then found the remotes package, which installed relatively quickly and downloaded webmorphR, but it choked compiling Rcpp. Maybe we just can't do this on a Raspberry Pi? (I love learning new things...) Hmmm... Still building webmorphR, and getting lots of reports like: /usr/include/c++/10/bits/vector.tcc:426:7: note: parameter passing for argument of type ‘std::vector<Catch::SectionEndInfo>::iterator’ changed in GCC 7.1 426 | vector<_Tp, _Alloc>:: | ^~~~~~~~~~~~~~~~~~~ In file included from /usr/include/c++/10/vector:67, from testthat/vendor/catch.h:671, from testthat/testthat.h:65, from test-runner.cpp:8: Do those mean this isn't going to work even if it does eventually finish? Does it require some specific version of GCC that isn't automatically selected? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.

[LorenAmelang]

My Pi is an old 32-bit version, but with the latest Debian Buster (Raspbian). Lots of things claim they run on it, but they don't. Probably more chance with a 64-bit Pi.

So I went to the other extreme from the Raspberry Pi, and tried installing on my serious Arch Linux box. Under ten minutes each for R and devtools, and less for webmorphR. No obvious problems, everything finished without error, and the only warnings were:

stri_container_listraw.cpp:63:41: warning: argument 1 value ‘18446744073709551615’ exceeds maximum object size 9223372036854775807

And:

In file included from src/uv-common.h:42,
                 from src/uv-common.c:23:
src/uv-common.c: In function ‘uv_walk’:
src/queue.h:68:19: warning: storing the address of local variable ‘queue’ in ‘MEM[(void *[2] * *)_2 + 8B]’ [-Wdangling-pointer=]

(There are a lot more lines of each if you're interested.)

Looks like your "hello world" type test uses your "demo stimulus files or functions that require python and dlib..."

I installed dlib, but R couldn't find a python package - seems odd. I tried install.packages("reticulate") which supposedly links R to any other python on my system. But I still get:

> devtools::install_github("debruine/webmorphR.dlib")
Downloading GitHub repo debruine/webmorphR.dlib@HEAD
Error in utils::download.file(url, path, method = method, quiet = quiet,  :
  download from 'https://api.github.com/repos/debruine/webmorphR.dlib/tarball/HEAD' failed
> devtools::install_github("debruine/webmorphR.stim")
Downloading GitHub repo debruine/webmorphR.stim@HEAD
Error in utils::download.file(url, path, method = method, quiet = quiet,  :
  download from 'https://api.github.com/repos/debruine/webmorphR.stim/tarball/HEAD' failed

Maybe if it wasn't "quiet" it wwould show what happens to the download? Can I turn that off somewhere?

I do have imagemagick-7.1.0.35-1 from the Arch repo. It says, "(2/2) Warn about old perl modules"...

Not your responsibility to teach me R, but I'm curious if any suggestions pop into your mind.