[Rd] compiling R under cygwin

[Latchezar (Lucho) Dimitrov]

> -----Original Message-----
> From: Prof Brian Ripley [mailto:ripley at stats.ox.ac.uk] 
> Sent: Friday, August 24, 2007 2:00 AM
> To: Duncan Murdoch
> Cc: Latchezar (Lucho) Dimitrov; r-devel at r-project.org
> Subject: Re: [Rd] compiling R under cygwin
> 
> On Thu, 23 Aug 2007, Duncan Murdoch wrote:
> 
> > The issue is compatibility with other Windows programs.  
> /path/to/file 
> > works in lots of Windows programs, and is interpreted 
> relative to the 
> > current drive.  In the common situation where the user only has one 
> > partition which is mounted as C:, it works (as long as they didn't 
> > switch to a CD or USB drive).
> 
> In particular, it does not work on my system where I need to 
> run under two OSes, Vista64 on C: and WinXP on D:, with 
> common development on E: .
> 
> I had originally intended to work entirely on E: and mount 
> E:/ as /, which is what you need to get Cygwin to return 
> /path/to/file for working files on E: .

There are many other options. Please check my e-mail with dir and ls
output.

> But Cygwin does not 
> handle permissions on a Vista NTFS file system well enough.

Would you please elaborate on this?

This one (as many others) I may consider Vista issues as well.

Again, I'm not pro or con cygwin. I actually do not want R on it.

> 
> Latchezar (Lucho) Dimitrov wrote
> 
> > 5. AFAIK 64-bit R on your favorite Windows is far far away (even 
> > Windows is not close enough although it's getting there) 
> and I'd like 
> > to see those above (and enjoy watching them) when Windows is not 
> > enough, e.g., microarray analysis/BioConductor.
> 
> That's just not true.  There have been reports of 64-bit R 
> ports to Windows and many people are running 64-bit Windows 
> on servers and some on desktops (including me).  A 
> MinGW-based 64-bit R for Windows is under active development 
> (which is why I have Vista64 installed

I suffer it too :-) and wholeheartedly sympathize you.

).

Thanks for the good news. Although it might be a little bit late - it
was not there when I badly needed it now that I have Solaris 10 OS x64 R
I barely would use it. Sorry. Please no bad feelings.

>  This is even 
> covered in the rw-FAQ (Q8.1, and yes that is a file with 
> numbered TOC).

OK, I take my words back - it's not "far far" it's only "far away" (or
even only "away" if this makes you happier :-). Meaning not _only R
itself but Windows as well_, i.e., the platform plus R. I admit my
initial statement was not that clear, sorry. As you can see here no
blames on R (core :-).

> 
> What seems 'far, far away' is something like Cygwin for 
> 64-bit Windows.

I agree but would use stronger expression here close to "never". And
this is one reason to try minGW (AFAIK they are close to have 64-bit of
it)

> (I saw on a Cygwin list that RedHat wanted $$$$$ to do it, 
> for too many $$$$$ for the potential customers.)  The 32-bit 
> version is said to work under the WOW 32-bit emulation layer 
> but many of us find it flaky.
> 
> Personally I use Linux (the main OS on the machine I mention 
> above) and
> Solaris: however many R users (quite possibly the substantial 
> majority of all R users even excluding student usage) want a 
> native Windows version.

I personally use whatever is most appropriate and hate it to be any
flavor of Windows but ... Again, how is Rgui more "native Windows
version" then plain native Windows CLI version? Are you using/confusing
"native"s two possible meanings here? One being minGW meaning of
"native" and the other as in "native Windows user" :-)))

Latchezar

> 
> -- 
> Brian D. Ripley,                  ripley at 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, UK                Fax:  +44 1865 272595
>