[R-pkg-devel] building binary gives error when trying to load

[Duncan Murdoch]

On 01/11/2016 1:10 PM, Henric Winell wrote:
> On 2016-11-01 17:01, Uwe Ligges wrote:
>
> > On 01.11.2016 15:56, Duncan Murdoch wrote:
> >> On 01/11/2016 9:49 AM, Rampal Etienne wrote:
> >>> Dear all,
> >>>
> >>> When I try to build a Windows binary in RStudio using R-devel,
> >>> everything goes well until the packages is tested whether it can be
> >>> loaded. I then get the error:
> >>>
> >>> ** testing if installed package can be loaded *** arch - i386 Error:
> >>> '\U' used without hex digits in character string starting "'C:\U"
> >>> Execution halted *** arch - x64 Error: '\U' used without hex digits in
> >>> character string starting "'C:\U" Execution halted ERROR: loading failed
> >>> for 'i386', 'x64' * removing
> >>> 'C:\Users\rampa\AppData\Local\Temp\Rtmp0o1Fm1\temp_libpath1cd445c31688/SADISA'
> >>>
> >>>
> >>> Error: Command failed (1) Execution halted Exited with status 1.
> >>
> >> That looks like a bug.
> >
> >
> > Indeed, it is also fairly new, as I typically do manual stuff that way
> > and it hit me 5 minuites ago. Sorry for not having tested it as I did
> > not believe there was a newly introduced bug.
> >
> > Uwe
>
> I reported this issue to R-devel about ten days ago.
>
> https://stat.ethz.ch/pipermail/r-devel/2016-October/073286.html

Yes, it's been there for a few weeks.  I suspect Uwe has a more complete 
set of environment variables than most people, and some of those could 
have masked the error.

As for me, I rarely develop packages on Windows any more.

Now that Microsoft is pushing their R products so much, you'd think they 
could contribute some time to supporting their platform.

Duncan Murdoch
>
>
> Henric Winell
>
>
>
> >
>
> >
> >
> >   It's in R, not RStudio.  To reproduce, start
> >> cmd.exe, and switch to a directory containing a source tarball or a
> >> source directory, and run
> >>
> >> R CMD INSTALL --build <tarball>
> >>
> >> For instance, when I do that with the tables package, I get
> >>
> >> D:\svn\MyR>R CMD INSTALL --build tables
> >> * installing to library 'f:\cygwin\home\murdoch\R\win-library\3.4'
> >> * installing *source* package 'tables' ...
> >> ** R
> >> ** preparing package for lazy loading
> >> ** help
> >> *** installing help indices
> >> ** building package indices
> >> ** installing vignettes
> >> ** testing if installed package can be loaded
> >> Error: '\c' is an unrecognized escape in character string starting
> >> "'f:\c"
> >> Execution halted
> >> ERROR: loading failed
> >> * removing 'f:\cygwin\home\murdoch\R\win-library\3.4/tables'
> >> * restoring previous 'f:\cygwin\home\murdoch\R\win-library\3.4/tables'
> >>
> >> It appears that the script is building paths including unescaped
> >> backslashes when it constructs the library path.  You can probably work
> >> around it by specifying the library path explicitly.  Since you were
> >> installing into a temporary library, that's a little tricky, but for
> >> most people it should be easy.
> >>
> >> I'll see if I can track it down and fix it.
> >>
> >> Duncan Murdoch
> >>
> >>>
> >>>
> >>> I have updated RStudio, Rtools and Rdevel,and the problem persists. Does
> >>> anyone have a clue how to solve this?
> >>>
> >>> Before I updated to the newest Rdevel, everything worked fine.
> >>>
> >>> Kind regards, Rampal Etienne
> >>>
> >>>     [[alternative HTML version deleted]]
> >>>
> >>> ______________________________________________
> >>> R-package-devel at r-project.org mailing list
> >>> https://stat.ethz.ch/mailman/listinfo/r-package-devel
> >>
> >> ______________________________________________
> >> R-package-devel at r-project.org mailing list
> >> https://stat.ethz.ch/mailman/listinfo/r-package-devel
> >
> > ______________________________________________
> > R-package-devel at r-project.org mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-package-devel
> >
>