Skip to content

R2.10 vs OSX 10.4.11

5 messages · Carl Witthoft, Simon Urbanek

#
Hi,
Can someone let me know what should work and what is known not to work 
with the latest R2.10x when installed on Tiger (10.4.11)?  I did a quick 
tryout and found that the ? command brings up a blank help window.  The 
help.start() (I think that's the command) brings up a R-help window in 
my web browser, but only some items' help pages show up there.

I'm happy to do some testing or debug work, but just don't want to beat 
myself up against known walls.

To answer the obvious:  I'm lucky still to have a Mac at work at all, so 
the chances of IT getting me a Leopard upgrade disc are zero or less.

thanks
Carl
#
On Jan 12, 2010, at 17:51 , Carl Witthoft wrote:

            
No, there is no "R2.10x".
R 2.10.1 was released for Mac OS X 10.4.11 and is the last release.  
There is no functional difference between the Leopard and Tiger  
release (other than the presence/absence of certain architectures), so  
upgrading the OS makes no difference in usability (for now -- there is  
likely to be no R 2.11.0 binary for Tiger anymore).

If you have any issues, please make sure you're using R 2.10.1 and you  
may properly report them.

Cheers,
Simon
#
Sorry if my shorthand (common enough among business software devs, of 
which I am not one) is misleading.  I simply meant 'the latest release 
of R2.10' when I wrote 2.10x.

In any case,  2.10 is working fine on my 10.5.8 OS at home, but the help 
window was not working on 10.4.11 at work.  I will post as much detail 
as I can acquire in the next day or two.  Given that R wasn't crashing, 
I may not have much to offer other than a list of which packages' help 
pages show up and which don't.

Carl
Simon Urbanek wrote:
#
Carl,
On Jan 12, 2010, at 8:07 PM, Carl Witthoft wrote:

            
Argh - please do include the *real* version! It *does* make a huge difference! For example 2.10.0 has some help-related bugs (and others ;)) that were fixed in 2.10.1 so it is important to distinguish exactly which version you have.
The new help system uses an integrated web server so firewall settings may get in your way there. I'm saying this only because you mentioned work ;).
A reproducible example is sufficient ... However, as I said there is no difference in the R help system between 10.4 and 10.5 so the reason is likely somewhere else (e.g. failure to re-install packages after an R upgrade).

Cheers,
Simon
#
I confirmed that I'm running OSX 10.4.11, on an Intel iMac (4 yrs old or 
so).  I am using Default FolderX, FinderPop, and Logitech COntrol Center 
Panes.

OK, here's the startup info from the console:

the console says..



*********
R version 2.10.1 (2009-12-14)
Copyright (C) 2009 The R Foundation for Statistical Computing
ISBN 3-900051-07-0
[skip dull stuff]

[R.app GUI 1.31 (5538) i386-apple-darwin8.11.1]

[Workspace restored from /Users/carl/Learning/Rgames/.RData]

Rgames: ?assign
starting httpd help server ... done
Rgames:

(where "Rgames" is my prompt.

AFter that,
help window opens, blank, etc.  If I use the "fuzzy match" window  the
package list shows up but nothing writes to lower half of window.

And as you asked, I am behind a firewall at work.  This is not something
I can modify.  So, how do I 'tell'  R not to go looking for online help?
I hope I didn't misunderstand your comment here -- but certainly no 
application should ever depend on an internet connection to function!

If I run help.start(), a page does open in my browser and I can search 
all packages which are on my machine.

But just typing ?assign,  or ? followed by any command from any basic 
package, or any package I've installed manually, will bring up the GUI's 
help window, which is always completely blank.  And I checked: menu 
commands like "select all" are dimmed, which is consistent with an empty 
document window.

I turned off all my Unsanity APE-related stuff, uninstalled 
ApplicationEnhancer, ran Disk Utility (no errors), and rebooted.  I also 
logged into my "emergency account," which I keep clean and unused, and R 
behaved the same way there.

Thanks for helping, and please let me know what other tests I can do to 
try to track down this problem.
Simon Urbanek wrote: