[R] issue building my own package... moving from Apple OS to Windows

Daryl Morris darylm at u.washington.edu
Sat Apr 25 02:28:54 CEST 2009


Thanks Mark,
That was it.  (ie the --binary option on R CMD build)

I'll just state for the web-o-sphere (just in case it helps any one 
else) that part of my confusion in building this came from 
misunderstanding what they call "source" and what the call "binary" in 
this environment.

I always think of source as something that must be compiled into a 
binary (like C-code for instance).  And I guess I didn't think the 
process of making some R functions into a loadable package was 
"compiling".  I thought it was more like "wrapping".  So, when the 
documentation referred to "source packages", I didn't think it was 
talking about what I was doing.  Nor did I think I had a "binary package".

Live and learn I guess.

It's still weird to me that I have to install "source packages" from the 
command line.  But I'm through the learning curve now, so I'll pipe down.

Thanks to everyone who responded!

Daryl Morris
Univ. of Washington Biostatistics



markleeds at verizon.net wrote:

Hii Daryl:  to install from the GUI or just generally on windows, build 
it so that it's by binary by doing R CMD BUILD --binary whatever.

that will create a zip file that you can either extract using winzip or 
install from the gui.
john chambers new  book explains some of this but, i  still found it 
difficult ( and got help also )  so
I know where you're coming from.

if you want relevant  package tutorials: duncan murdoch ,, friedrich 
leisch and alan lenarcic all have them on the web.  but i think you are 
way past the point where you need them. ( and I couldn't
get them to work anyway but they were still quite helpful ).

as far as not needing Perl, I don't know how to do that ? I think they 
are changing the CMD
commands to use R rather than Perl so, when that is finished, you won't 
need Perl.

oh, if you do do any of those tutorials and get any of them to woprk, 
let me know because
I couldn't. thanks.




More information about the R-help mailing list