[Bioperl-l] Installing OS X dependencies

[Steven Cannon]

Todd:

Thanks for the t-coffee installation suggestion. I had success this 
morning with a similar fix, suggested by Bill Van Etten, who wrote:

The hack/solution I came up with is to comment out the declaration and 
implementation of vfree in the t_coffee source such that the vfree in 
libm is used whener t_coffee calls it. Only required commenting out a 
couple lines in two files (t_coffee_source/util_declare.c and 
t_coffee_source/util_lib_header.h). The diffs are provided below. I'm 
guessing that this hack works as the tests (performed during the build) 
reported 100% success.

This seemed to work. I did the same sort of thing with Tree-Puzzle for 
OS X (a phylogenetic program), which also had a namespace collision. The 
change was to rename the occurances
of radixsort in ml.h and ml1.c to tpradixsort.

I would be interested in seeing some sort of common repository of these 
kinds of 'make' fixes for OS X (particularly for the Bioperl 
dependencies, but also for related bits of bioinformatic software).

Steve



On Monday, July 2, 2001, at 11:43 AM, Todd Richmond wrote:

> On 6/30/01 12:34 PM, "Steven Cannon" <cann0010@tc.umn.edu> wrote:
>
>> Bioperl 0.7.1 installed nicely on my Mac OS X, but I have found that
>> installing the dependencies is a challenge (particularly for me as a
>> *nix novice). I have found OS X binaries or installs for Clustalw, 
>> ncbi,
>> Xfree86, and Lesstif (open source Motif clone, which has libraries the
>> ncbi Toolkit may use?) But I have not been able to install t-coffee. I
>> suspect that only small changes need to be made to the makefiles, but I
>> don't know what those changes should be to get the install working on
>> Darwin. Advice?
>>
>
> Here's what my "install" file looks like:
>
> -------------------------------------
> cd t_coffee_source
> make clean
> make -i CC="cc -Dvfree=my_vfree" USER_BIN=../bin/ all
> cd ..
> cd ./lalign
> make clean
> make -i CC=cc USER_BIN=../bin/ all
> cd ..
> ./bin/test.pl
>
> -------------------------------------
>
> The important change here is the "-Dvfree=my_vfree" portion. If you 
> don't
> redefine the vfree variable for the compiler, make complains about 
> symbols
> that are defined multiple times (if I remember correctly - I installed 
> it
> awhile ago). Let me know if this doesn't work, and I'll try to figure 
> out if
> I had to do something else as well.
>
>
> --
> Todd Richmond                    http://cellwall.stanford.edu/todd
> Carnegie Institution             email: todd@andrew2.stanford.edu
> Department of Plant Biology      fax: 1-650-325-6857
> 260 Panama Street                phone: 1-650-325-1521 x431
> Stanford, CA 94305
>
>