[Rd] problem building R 2.12.1 64-bit on Windows 7

Michael Sumner mdsumner at gmail.com
Fri Dec 31 05:12:01 CET 2010


It seems the problem was as elementary as me mis-interpreting the
instructions in 3.1.3 of the R Installation and Administration manual:

"Look at ‘MkRules.dist’ and if settings need to be altered, copy it to
‘MkRules.local’ and edit the settings there."

I took that to mean that you only copy the settings that need to be
altered to a new local file, so the majority of required settings
weren't passed to MkRules itself. "Copy it" means copy the file, not
just the setting/s - a poor reading on my part, and easily checked by
seeing the resulting MkRules file.

make all recommended ran to completion without error, and with
Kaspersky in full operation.

Thanks again for the patience, I've added my updated notes below.

Cheers, Mike.


Using Rtools212.exe, with MiKTeX, no static HTML help, and no
installer build.

Everything installed is Run As Administrator, including the command
prompt used to build R.

Environment Variables

These environment variables are set:


Rtools and MiKTeX installed, with the following at the start of PATH:



Options to avoid messages when extracting sources:

--no-same-owner --no-same-permissions


I can write to this directory:


Build steps

1) Extract from the archive

tar -xf R-patched_2010-12-27_r53886.tar.gz

2) Copy Rsrc "src/" and "Tcl/" into R-patched/ from the relevant 64
bit Rtools source

3) Copy [R_HOME]/src/gnuwin32/MkRules.dist as MkRules.local and make
the following changes (the jpeg setting is not required I think, I
just want to report exactly what I did)

diff MkRules.dist MkRules.local

< WIN = 32
> WIN = 64
< # JPEGDIR = jpeg-8a
> JPEGDIR = jpeg-8a

4) Make

cd R-patched/src/gnuwin32
make all recommended

On Fri, Dec 31, 2010 at 10:53 AM, Michael Sumner <mdsumner at gmail.com> wrote:
> Hello, thank you both - in an earlier test I turned off Kaspersky
> completely to check that but saw no difference, but I'm only going on
> my memory for that.
> I'll check again more completely later today and also try turning on
> the reporting for the sed line in MkRules as Duncan suggested.
> I don't have an override for SYMPAT or SYMPAT64.
> Cheers, Mike.
> On Fri, Dec 31, 2010 at 10:34 AM, peter dalgaard <pdalgd at gmail.com> wrote:
>> On Dec 30, 2010, at 23:19 , Duncan Murdoch wrote:
>>> On 10-12-30 4:13 PM, Michael Sumner wrote:
>>>> Hello,
>>>> ...
>>>> make[4]: *** [Rzlib.dll] Error 4
>>>> make[3]: *** [rlibs] Error 1
>>>> make[2]: *** [../../bin/x64/R.dll] Error 2
>>>> make[1]: *** [rbuild] Error 2
>>>> make: *** [all] Error 2
>>> I don't know what's going wrong.  At that point it is supposed to be making Rzlib.dll (as the error at the end shows it was).  That is made using src/extra/zlib/Makefile.win. I don't see any calls to "sed" in there, but there's one in the implicit rule to make the .dll (in src/gnuwin32/MkRules), namely
>>> $(SED) -n $(SYMPAT)
>>> Do you have a definition for SYMPAT that overrides ours, or SYMPAT64 (which is used in ours)?
>>> If that's not it, you could remove the @ sign from the beginning of that line in MkRules, and see what it is trying to do just before it dies.
>>> Duncan Murdoch
>> Off-the-cuff: Is there a virus scanner active on the system? We have had a couple of reports that turned out to be antivirus software swiping files away for checking right under the nose of their confused maker...
>> --
>> Peter Dalgaard
>> Center for Statistics, Copenhagen Business School
>> Solbjerg Plads 3, 2000 Frederiksberg, Denmark
>> Phone: (+45)38153501
>> Email: pd.mes at cbs.dk  Priv: PDalgd at gmail.com
> --
> Michael Sumner
> Institute for Marine and Antarctic Studies, University of Tasmania
> Hobart, Australia
> e-mail: mdsumner at gmail.com

Michael Sumner
Institute for Marine and Antarctic Studies, University of Tasmania
Hobart, Australia
e-mail: mdsumner at gmail.com

More information about the R-devel mailing list