Skip to content

Spinning wheel when running script

7 messages · Rick Johnson, Simon Urbanek, David Winsemius +1 more

#
On Jul 28, 2011, at 8:32 AM, Rick Johnson wrote:

            
Spinning wheel means that R has entered C code that is not interruptible. It's no different from the Windows version except that Windows GUI doesn't even have the indicator.

If you are running a script (interpreted code), you can interrupt it with <Esc> or the Stop button and there is no spinning wheel.

Cheers,
Simon
#
On Jul 28, 2011, at 8:32 AM, Rick Johnson wrote:

            
There are various reasons:

1) Big job  .. takes time. 5-10 minutes for some of mine. I even have  
ones that take 45 minutes to an hour.
2) variant of 1) Physical memory  exhausted, virtual memory being  
used, performance becomes 10-20 times slower.
In that instance it is useful to open the Activity Monitor.app in  
your /Applications/Utilities folder. I keep it on my Dock, where it  
displays memory availablity. I then know if virtual memory is getting  
accessed and can then go get another cup of coffee.
3) Program bug, enters loop, lost connection with your OS, refuses to  
respond to <esc> or Stop button,  needs to have "Force Quit" from the  
Finder.
Most of the cognoscenti seem to use emacs and Sweave/LaTex, although  
there are other smart  people using Eclipse and an increasing number  
using RStudio. Both of the first failed , in the first case (emacs)  
due to my inability to adapt to the new keystrokes and getting stuck  
in buffers that I could not exit,  and in the case of Eclise/StatET  
due to inadequate directions to support installation. (I do read  
manuals and follow directions reasonably well but they need to have  
_all_ the steps and at least when StatET was being rolled out, there  
were definitely missing steps. I am a Unix-weenie, so some steps that  
would seem second nature to the *Nixerati are not in my repertoire. It  
is useful to have some Unix commands like locate, ls, and cd, and it  
is very useful to discover that dragging files to the Terminal.app  
command line will auto-compose a full file-path spec. )  Rstudio looks  
nice except Simon and Hans-Peter have continued to improve the R-GUI  
and I like it.  I regularly save a console transcript, edit out the  
errors and also same my useful work to text files or to spreadsheet.  
The RGUI for the Mac allows  you to do your editing in separate files  
now with an auto-save option that is very handy. You can select and  
hit cmd-return and get transfer to and immediate execution in the  
Console.

Now, if I do need to force quit, at least I recover my code that has  
been saved in another place.
#
On 28-07-2011, at 14:32, Rick Johnson wrote:

            
I use TextMate ( http://macromates.com/ ) and the R bundles developed by Hans-J?rg Bibiko a lot.
Very nice.

Berend
#
Thank you David and Simon for quick responses. I was hoping that I might have made some usage error that you had seen before.  looks like I will have to dig deeper.

This is not a big job - yet.  I am just starting with simple intro lines:  library function, setting some file path names, and setting color scheme variables.  Code that works fine in my windows R project. No loops in my code, but there might be in the package functions. Not the point when I would go for a cup of coffee.

Plenty of memory.  Activity monitor shows maybe 63MB for R64 app and over 5 G free. R64 uses 45% of cpu.

Esc and R console stop button do not interrupt the process - consistent with  what Simon wrote: "Spinning wheel means that R has entered C code that is not interruptible."  The activity monitor says "not responding".  I have to kill it with the activity monitor.  Then the system produces a forced quit problem details report which I can mail if that is of any use.  It describes the event as "hang".

I am using the text editor that comes with the r.app gui.  I wasn't aware of Rstudio and will see if I have better success with that. 

My intuition is that the r app gui is having problems with multi line paste of script, but your comments point to some error in the user code.

My script is simple, but it occurs to me that one of the packages might be a potential problem source.  I'll try testing them one at a time and see if that pinpoints a problem. And, I'll let you know if I make progress.  thanks

Rick
On Jul 28, 2011, at 7:17 AM, David Winsemius wrote:

            
#
Hello,

I should have loaded RStudio before responding.  My intro code runs fine within RStudio. 
Now I can get to the cup of coffee stage.

Thanks Thanks

Rick Johnson

Inverness, CA
#
On Jul 28, 2011, at 1:04 PM, Rick Johnson wrote:

            
We would appreciate if you could pinpoint the issue, though, so if it's really R.app's fault, we'd like to fix it ...
Can you send me the script in question or something like that?

Thanks,
Simon