[rt-users] Perl conflicts

Tim Cutts tjrc at sanger.ac.uk
Wed Oct 31 06:10:56 EDT 2012

On 30 Oct 2012, at 19:47, Thomas Sibley <trs at bestpractical.com> wrote:

> On 10/30/2012 12:21 PM, Ram wrote:
>> How do folks deal with perl conflicts? We normally use RPMs for
>> everything but that's not practical given the relatively high-version
>> requirements of the rt4 branch so CPAN seems the only practical
>> approach.
> Use a completely separate build of perl just for RT instead of the
> system installed and managed perl.
> Look at perlbrew for an easy way to build perl if you're not familiar
> with the process.

I agree that's probably the most pain-free and robust method.

On my Mac, where I do some RT tinkering as a standalone build, I have an rt-support directory in my home directory, and I put all the libraries RT needs in there, and I point my PERL5LIB and DYLD_LIBRARY_PATH variables at it whenever I'm working on RT.  The relevant bits of my CPAN/MyConfig.pm file:

  'makepl_arg' => q[INSTALL_BASE=~/rt-support],
  'mbuildpl_arg' => q[--install_base ~/rt-support],

and then my PERL5LIB is:


On my production RT server, which is running Ubuntu, I use the system perl, and install pre-packaged modules where I can, but if I can't I just let RT's 'make fixdeps' do what it likes, and install the necessary packages in the system's site perl directory.  Since the server isn't used for anything other than RT, I'm not bothered about superseding what the OS itself installs.  The major advantage of doing it that way is that it's then much easier to configure with the packaged versions of apache, mod_perl and so on.


 The Wellcome Trust Sanger Institute is operated by Genome Research 
 Limited, a charity registered in England with number 1021457 and a 
 company registered in England with number 2742969, whose registered 
 office is 215 Euston Road, London, NW1 2BE. 

More information about the rt-users mailing list