[R-SIG-Mac] Re: [Rd] Re: Double-clicking Darwin-Aqua R-devel

[Simon Urbanek]

On Friday, June 13, 2003, at 12:55  PM, G. Sawitzki wrote:

> Stefano, thank you for the Darwin-Aqua R-devel.
Yes, I think it's the first step in the right direction - we need a 
native OS X GUI ;). Good job, Stefano.

> here are first observations:
>
> 1) the application must be named exactly as distributed. Usually I 
> would ad a version number, but this leads to
That should be easy to fix since there is a call to the framework to 
ask where the bundle root is.

> 2) libdl.0.dylib must be installed, eg. from fink. Else you get
> dyld: /Applications/RAqua.app/R.bin.app/Contents/MacOS/R.bin can't 
> open library: /usr/lib/libdl.0.dylib  (No such file or directory, 
> errno = 2)
A more general question about this: the libdl is really very tiny since 
all it does is it maps the usual "dfcn.h" API to Apple's API. Why not 
to add this statically to the R code? I'm not sure that this really 
needs to be external/dynamic code. Besides each version of the library 
differs (prepending underscores etc.) - so shouldn't we simply supply 
it with R in general (not just with the RAqua) like we do with pcre?

> 3) it seems two processes are lauched, and both show up in the dock:
> 17562  ??  S      0:00.01 /Applications/RAqua.app/Contents/MacOS/RAqua 
> -psn_0_6291457 HOME=/Users/gs SHELL=/bin/tcsh ....
> 17563  ??  S      0:09.72 
> /Applications/RAqua.app/R.bin.app/Contents/MacOS/R.bin --gui-aqua 
> PWD=/ R_HOME=/Applications/RAqua.app/R.bin.app/Contents 
> R_SESSION_TMPDIR=/tmp/Rtmp17563 ....
Obviously one is the GUI and the other the "regular" R (by using just 
one integrated bundle this problem should disappear). Funny enough one 
the the two Rs "jumps" in the dock up and down all the time until it's 
tired ;). A rather technical question: How do the two processes 
communicate?

4) The GUI should set the option to make "quartz" the default device 
(so far options(device="quartz") works quite well).

Cheers,
Simon

---
Simon Urbanek
Department of computer oriented statistics and data analysis
University of Augsburg
Universitätsstr. 14
86135 Augsburg
Germany

Tel: +49-821-598-2236
Fax: +49-821-598-2200

Simon.Urbanek at Math.Uni-Augsburg.de
http://simon.urbanek.info