You could also preprend your cat so it outputs something like
INCLUDE_DIR=..
and then use grep and sed to trim it. Of course that can also end up being
brittle.
On Wed, Apr 7, 2021 at 10:48 PM Henrik Bengtsson <henrik.bengts...@gmail.com>
wrote:
> Correcting: the env var is R_LIBS_USER and nothing else.
>
> /Henrik
>
> On Wed, Apr 7, 2021 at 1:20 PM Henrik Bengtsson
> <henrik.bengts...@gmail.com> wrote:
> >
> > Can you go via a temporary file instead, i.e. output what you want to
> > grab to a temporary file and pull the info from that instead?
> >
> > Side tracking ...
> >
> > > I thought about using --vanilla but I use .Rprofile to set my library
> path,
> > > so ignoring that file completely defeats the point in my setup.
> >
> > I'm a proponent of customizing the library path via .Renviron instead
> > of via .Rprofile. In your case, you can have a line in ~/.Renviron
> > with:
> >
> > R_USER_LIBS=~/R/%p-library/%v-bioc_3.12
> >
> > Alternatively, you can set it elsewhere, e.g. ~/.bashrc, in a Linux
> > environment module that is loaded, and so on.
> >
> > BTW, using Rscript --no-init-file ... would skip .Rprofile while still
> > parsing .Renviron.
> >
> > /Henrik
> >
> > On Wed, Apr 7, 2021 at 8:16 AM Mike Smith <grimbo...@gmail.com> wrote:
> > >
> > > I have the following line in the configure.ac for rhdf5filters, which
> > > returns the location of the HDF5 headers distributed with Rhdf5lib:
> > >
> > > RHDF5_INCLUDE=`"${R_HOME}/bin${R_ARCH_BIN}/Rscript" -e
> > > 'cat(system.file("include", package="Rhdf5lib"))'`
> > >
> > > For me the output is a path like
> > > /mnt/data/R-lib/4.0.3-bioc_3.12/Rhdf5lib/include, which gets inserted
> into
> > > the package Makevars file, and the package compilation works.
> > >
> > > However I've had multiple reports (
> > > https://github.com/grimbough/rhdf5filters/issues/11) where this
> doesn't
> > > work, all of which seem to relate to messages printed when an Rprofile
> is
> > > loaded. They have well intentioned messages like below, which don't
> work
> > > so well when passed as compiler flags
> > >
> > > [1] "[BMRC] You have sourced the BMRC Rprofile provided at
> > > /apps/misc/R/bmrc-r-user-tools/Rprofile"
> > > [1] "[BMRC] Messages coming from this file (like this one) will be
> > > prefixed with [BMRC]"
> > > [1] "[BMRC] You are running R on host <rescomp3.hpc.in.bmrc.ox.ac.uk>
> > > with CPU <Intel(R) Xeon(R) CPU E5-2650 v2 @ 2.60GHz>"
> > > [1] "[BMRC] While running on this host, local R packages will be
> > > sourced from and installed to
> > > /well/combat/users/ifl143/R/4.0/ivybridge"
> > > /gpfs3/well/combat/users/ifl143/R/4.0/ivybridge/Rhdf5lib/include
> > >
> > > I thought about using --vanilla but I use .Rprofile to set my library
> path,
> > > so ignoring that file completely defeats the point in my setup. Is
> anyone
> > > aware of either a more reliable way of getting the information I want
> > > (maybe suppressing messages, different mechanism entirely, etc)?
> > > Alternatively, is there anything definitive in WRE or the like that
> > > suggests printing messages Rprofile is a bad idea that I can pass on
> to the
> > > users?
> > >
> > > Cheers,
> > > Mike
> > >
> > > [[alternative HTML version deleted]]
> > >
> > > _______________________________________________
> > > Bioc-devel@r-project.org mailing list
> > > https://stat.ethz.ch/mailman/listinfo/bioc-devel
>
> _______________________________________________
> Bioc-devel@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>
--
Best,
Kasper
[[alternative HTML version deleted]]
_______________________________________________
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel
