[BioC] Rgraphviz load problem

[Kasper Daniel Hansen]

Daniel

By now you may have a bit of a messed up installation of graphviz. As  
Li, I would like to know the following
(1) Have you installed the CRAN compilers? (look for the existence  
of /usr/local/gcc4.0)
(2) I assume you have the CRAN version of R - right?
(3) What is your default C compiler
  # gcc_select
(hopefully it is 4.0)
(4) How did you install graphviz
(5) And I would like the output of installing Rgraphviz under source  
(important).

You should not have to DYLD_LIBRARY_PATH, indeed it is a bit  
dangerous to do so under OS X (I am told).

If you want to clean up the whole mess, do the following
(a) delete all the graphviz versions you have. Perhaps using
# make uninstall
from you building directory. There are files both in /usr/local/bin, / 
usr/local/include, /usr/local/lib
(b) Install graphviz 2.12 by running configure and make. I have used  
the CRAN compilers to compile graphviz, just for safety, but I think  
you can just use the normal gcc (= don't do anything) _if_ you are  
using 4.x
(c) Download Rgraphviz from the development version of Bioconductor.
(d) Install it by
# R CMD INSTALL ....

That should work.

Kasper

On Apr 14, 2007, at 9:35 AM, hpages at fhcrc.org wrote:

> Hi Daniel, Seth,
>
> Quoting Daniel Klevebring <daniel at biotech.kth.se>:
>
>> After I installed graphviz 2.8 I can't run the command line tools of
>> graphviz:
>> danielk:~ danielk$ dot
>> dyld: Library not loaded: /usr/local/lib/graphviz/libgcc_s.1.0.dylib
>>    Referenced from: /usr/local/lib/graphviz/libgvc.2.dylib
>>    Reason: image not found
>> Trace/BPT trap
>
> The graphviz-2.8.dmg available at http://r.research.att.com/
> has been broken for months:
>   https://stat.ethz.ch/pipermail/r-sig-mac/2007-January/003534.html
>
> You can "fix" it by adding a symlink to
>   /usr/local/gcc4.0/lib/libgcc_s.1.0.dylib
> in /usr/local/lib/graphviz/
>
> (alternatively you can send a friendly reminder to R-SIG-Mac
> that graphviz-2.8.dmg is broken)
>
> Cheers,
> H.
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/ 
> gmane.science.biology.informatics.conductor