[ome-devel] OME install woes

Ilya Goldberg igg at nih.gov
Wed Apr 12 21:02:47 BST 2006


On Apr 11, 2006, at 4:25 PM, august wrote:

> My concern at the moment is whether or not an install script is the
> right way to tackle the problem of installation.   It seems to me that
> there is no real way to cover all the different modules and  
> dependancies
> that go with each and every system and setup.   I feel it would be
> better to have a "recipe" or "blueprint" for installation which is
> broken down to the basics :
> 	what you need the web server to do, what you need perl and
> 	mod_perl to do, and what you need postgres to do.

I think you're right.  The installer really needs better guides for  
pre-installation by defining how to set up and check the things you  
say.  Once that stuff is set up properly, the installer can then be  
run non-interactively, in which case it just performs a bunch of  
tests before installing OME.  This is essentially how the OS X  
installer GUI works.  It installs binary packages for all the  
dependencies, edits the text version of /etc/ome-install.store  
setting various system things, and launches the regular OME installer  
non-interactively, which only checks all that stuff once more before  
proceeding to install OME.

Good luck with CVS.  I think you'll find it much better.
The OS X installer can be got here (its ~20 M) in case you're  
interested:
http://users.openmicroscopy.org.uk/~igg/OME-Installer.dmg
The binaries distributed there will not work on intel macs.

-Ilya


> best -august.
>



More information about the ome-devel mailing list