GBrowseInstallation

From WormBaseWiki
Revision as of 07:45, 14 February 2008 by Smckay (talk | contribs)
Jump to navigationJump to search

Notes on installing Gbrowse in the WormBase environment.

One barrier to packaging GBrowse with the rest of the wormbase code-base in CVS has been the architecture dependent paths that GBrowse uses by default. The following recipe gets rid of this "feature"

1) Check out a fresh copy of the 'stable' branch, or run 'make realclean' and do a CVS update. Create or edit the file GGB.def in the same directory as Makefile.PL.

GGB.def

# This file is read by Makefile.PL

CGIBIN=/usr/local/wormbase/cgi-perl/seq
DO_XS=0
APACHE=/usr/local/apache
CONF=/usr/local/wormbase/conf
GBROWSE_ROOT=gbrowse
HTDOCS=/usr/local/wormbase/html
#LIB=/usr/local/wormbase/lib
VERSION=1.69
BIOGRAPHICS_VERSION=1.654

2) Run perl Makefile.PL with the usual options plus DO_XS=0. You can either edit the GGB.def or use the command line args. To contain the perl libraries, you can set the LIB option as well. The DO_XS option is why the GBrowse libs have been going into architecture-dependent paths. It concerns a tiny but rarely-used bit C code associated with one of the aligner plugins. It is not used in the wormbase.

 perl Makefile.pl

3) Do the installation with the incantation below.

 sudo make install UNINST=1

4) The UNINST=1 should remove the previous installation of libs in the architecture dependent paths but you will want to check to be sure. Manually prune the directory /system/dependent/path/Bio if necessary. Check first to make sure bioperl is not installed in there (it should not be).