On Feb 18, 2009, at 7:54 AM, Ben Webb wrote:
Daniel Russel wrote:I think we need to put some thought towards what ways of using IMP we want to support. As I see it, there are two usage models to support:1) The user installs IMP in some standard location and uses it like anyother library....2) The user adds a module to IMP (but does not check it in to IMP svn)and uses the IMP build system and imppy.sh to run things.(1) is pretty much the standard way "normal" packages work, so I think should be the only one we support. We could possibly provide a simple set of SConscripts for those wanting to link programs or other modules against IMP.
My reasons for suggesting have 2 around are:- it is nice to be able to use IMP in place when developing since you typically have multiple copies of the library that you are using (I would strongly encourage people who are checking things in to have a copy which is the current svn version so you can test your checkins easily for missed files and stuff) - it is a pain to have to propagate changes back from installed headers to SVN ones (the debugger for external programs will take you to installed headers and it is easy to get confused about which version of the headers you are editing). - not everyone around seems to be comfortable setting up a build system, so it is nice to have a canned solution for them. - we already have all the necessary hacks in place since that is how you set things up initially :-)