[Rd] arbitrary size data frame or other stcucts, curious about issues invovled.

2011-06-12 Thread Mike Marchywka

We keep getting questions on r-help about memory limits  and
I was curious to know what issues are involved in making 
common classes like dataframe work with disk and intelligent
swapping? That is, sure you can always rely on OS for VM 
but in theory it should be possible to make a data structure
that somehow knows what pieces you will access next and
can keep thos somewhere fast. Now of course algorithms 
should act locally and be block oriented but in any case
could communicate with data structures on upcoming 
access patterns, see a few ms into the future and have the 
right stuff prefetched.
 
I think things like bigmemory exist but perhaps one
issue was that this could not just drop in for data.frame
or does it already solve all the problems?
 
Is memory management just a non-issue or is there something
that needs to be done  to make large data structures work well?
 
 
Thanks.
 
 
---

415-264-8477
marchy...@phluant.com

Online Advertising and Analytics for Mobile
http://www.phluant.com
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


Re: [Rd] Require of gWidgetsRGtk2 fails: RGtk2.dll can't be found, but it's there

2011-06-12 Thread Tony Plate
I also had frustrations trying to use libary RGtk2 under Windows XP.

I kept getting the message
   unable to load shared object C:\R\site-library\RGtk2\libs\i386\RGtk2.dll
even though it was on my path.

Uninstalling and reinstalling various versions of GTK2, both through R
and outside R, many times, including rebooting Windows, did not help.

With the help of some comments on R-help (Neil Rice, Mar 7, 2011), I found that
there was another conflicting version of zlib1.dll in a directory on my PATH.
Removing that directory from the path (inside R) fixed the problem and I am able
to run RGtk2 now.

Here's an R expression to look for copies of the various DLLs that GTK2 uses 
(change c:/GTK2-Runtime/bin/ to be appropriate for your system):

  for (dll in list.files(c:/GTK2-Runtime/bin/, pattern=*.dll$)) 
  print(with(list(x=file.path(strsplit(Sys.getenv(PATH), ;)[[1]], dll)), 
  x[file.exists(x)]))
[1] C:\\GTK2-Runtime\\bin/freetype6.dll
[1] C:\\GTK2-Runtime\\bin/intl.dll
[1] C:\\GTK2-Runtime\\bin/libatk-1.0-0.dll
[1] C:\\GTK2-Runtime\\bin/libcairo-2.dll
[1] C:\\GTK2-Runtime\\bin/libcairo-gobject-2.dll
[1] C:\\GTK2-Runtime\\bin/libcairo-script-interpreter-2.dll
[1] C:\\GTK2-Runtime\\bin/libexpat-1.dll
[1] C:\\GTK2-Runtime\\bin/libfontconfig-1.dll
[1] C:\\GTK2-Runtime\\bin/libgailutil-18.dll
[1] C:\\GTK2-Runtime\\bin/libgdk-win32-2.0-0.dll
[1] C:\\GTK2-Runtime\\bin/libgdk_pixbuf-2.0-0.dll
[1] C:\\GTK2-Runtime\\bin/libgio-2.0-0.dll
[1] C:\\GTK2-Runtime\\bin/libglib-2.0-0.dll
[1] C:\\GTK2-Runtime\\bin/libgmodule-2.0-0.dll
[1] C:\\GTK2-Runtime\\bin/libgobject-2.0-0.dll
[1] C:\\GTK2-Runtime\\bin/libgthread-2.0-0.dll
[1] C:\\GTK2-Runtime\\bin/libgtk-win32-2.0-0.dll
[1] C:\\GTK2-Runtime\\bin/libpango-1.0-0.dll
[1] C:\\GTK2-Runtime\\bin/libpangocairo-1.0-0.dll
[1] C:\\GTK2-Runtime\\bin/libpangoft2-1.0-0.dll
[1] C:\\GTK2-Runtime\\bin/libpangowin32-1.0-0.dll
[1] C:\\GTK2-Runtime\\bin/libpng14-14.dll
[1] C:\\GTK2-Runtime\\bin/zlib1.dll
 

If there are any lines beginning with [2] in the output, there are multiple
versions of DLLs and these are potential problems (check that either the
DLLs are the same version, or that the GTK2 version is ahead on the path).

Note that is is not sufficient that the GTK2 library is on the path -- it has
to be ahead of other path components that have conflicting DLLs.


On 6/11/2011 8:59 PM, Michael Lawrence wrote:
 I'm working on a mechanism that will download GTK+ (the official zip files)
 into a predetermined location and put that location in front of the other
 paths at load time.  Hopefully this will resolve these continuing issues.

 Michael

 On Fri, Jun 10, 2011 at 6:07 AM, Janko Thyson
 janko.thyson.rst...@googlemail.com  wrote:

 Okay, at least some positive result today (language issue ;-))

 Thanks for trying. Seems like GTK2 and me are not meant to be ...

 Regards,
 Janko


 On 10.06.2011 14:36, Prof Brian Ripley wrote:

 It is a Microsoft error message, so your Windows is translating it, not R.
   It does say 'The specified module' in English Windows, without saying who
 specified the module (but it does often produce a popup naming the module).

 I've no better idea what your problem is: the @ReadMe instructions work
 for me, Uwe on winbuilder and in our teaching lab (and the advice in RGtk2
 does not).

 On Fri, 10 Jun 2011, Janko Thyson wrote:

   On 10.06.2011 13:18, Prof Brian Ripley wrote:
   On Fri, 10 Jun 2011, Janko Thyson wrote:

 Dear list,

 I've been trying to get gWidgets/gWidgetsRGtk2 to run
 every other month for a while, but somehow I simply can't
 figure out what's going wrong.


   Your subject line indicates your confusion.  It does not say
   RGtk2.dll cannot be found (at least, the English version of the
   message does not say so).

   What it means is that RGtk2.dll or one of the DLLs it depends on
   cannot be found.  See the instructions at
   http://cran.r-project.org/bin/windows/contrib/2.13/@ReadMe

   It is Microsoft's error message, not ours.


 Thanks for your answer. I followed the advice in the readme file and
 installed


 http://ftp.gnome.org/pub/gnome/binaries/win32/gtk+/2.22/gtk+-bundle_2.22.0-201010
 16_win32.zip
 Previously I made sure that all other previously installed GTK+ runtime
 environments were removed and that the windows PATH reflects the correct
 path to
 the new GTK+ files. I also made sure I started a new R session before
 trying
 'require(gWidgetsRGtk2)' again. Yet, the same error. More precisely, a
 dialog box
 pops up asking me to install GTK+ or not. That's what's confusing me as
 well: the
 package's recommendation with respect to the version of an GTK+ runtime
 environment is somewhat out-of-sync with the advice in the readme, even
 though it
 the version is fairly recent (2011-04-30)

 Sorry for the german part in the error message, but I wasn't able to get
 it all
 to English (tried language = en in Rconsole file, setting up a
 

[Rd] Cross-compiling package

2011-06-12 Thread Sylvain Loiseau
Hello,

I'm facing problems with compiling a package (containing only R-code and 
dataset) under OSX for Windows.

Is there a command line producing a zip archive that can be installed on 
Windows using the windows R GUI ?

Best,
Sylvain
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


Re: [Rd] Cross-compiling package

2011-06-12 Thread Uwe Ligges



On 12.06.2011 18:57, Sylvain Loiseau wrote:

Hello,

I'm facing problems with compiling a package (containing only R-code and 
dataset) under OSX for Windows.

Is there a command line producing a zip archive that can be installed on 
Windows using the windows R GUI ?


Well, cross compiling was always rather difficult.  Why not just you the 
winbuilder service, if you do not have access to a Windows box:

http://win-builder.r-project.org/
It also provides you with check results for the Windows platform.

Best,
Uwe Ligges




Best,
Sylvain
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


Re: [Rd] Cross-compiling package

2011-06-12 Thread Prof Brian Ripley

On Sun, 12 Jun 2011, Uwe Ligges wrote:


On 12.06.2011 18:57, Sylvain Loiseau wrote:

Hello,

I'm facing problems with compiling a package (containing only R-code and 
dataset) under OSX for Windows.


Packages are 'installed': compilation is only involved if they contain 
C/Fortran/C++/... source code.  It seems using the wrong term has 
misled people.


Is there a command line producing a zip archive that can be installed on 
Windows using the windows R GUI ?


Well, cross compiling was always rather difficult.  Why not just you the 
winbuilder service, if you do not have access to a Windows box:

http://win-builder.r-project.org/
It also provides you with check results for the Windows platform.


Good point, but there is no 'cross-compiling' involved here.

Suppose I have a package 'mypkg' installed on Mac OS X (to give it the 
correct name).  Then in a Terminal window


cd directory above mypkg
zip -r9X mypkg_ver.zip mypkg

where you need to set 'ver' to be the version number of the package, 
read from mypkg/DESCRIPTION.




Best,
Uwe Ligges




Best,
Sylvain
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel



--
Brian D. Ripley,  rip...@stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford, Tel:  +44 1865 272861 (self)
1 South Parks Road, +44 1865 272866 (PA)
Oxford OX1 3TG, UKFax:  +44 1865 272595

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel