Daniel- You've pretty much hit the high spots as far as I know. I had planned to set up some virtual box machines for such testing (and will eventually).
I haven't done it myself but if you put the linux on a VM, you could snapshot the VM with different sets of dependencies installed (for example). Then just pick the appropriate VM to start for the level of testing you desire. I suggest serial testing since after you work through the first flavor of OS and all the different PDL issues, the next will be *very* similar if not identical. The main reason for more comprehensive testing is so that the one per-platform issue (as in "oops, I forgot that XX needs YY and not ZZ") can be an insurmountable obstacle to a new PDL user.... OpenSolaris would be great but I don't think we have any recent tests with PDL at the moment. An issue there is the OpenGL + FreeGLUT dependencies since I don't think Solaris includes FreeGLUT by default. --Chris On 7/24/2010 3:22 PM, Daniel Carrera wrote: > This message is mainly for Chris. > > If it helps, I volunteer to test the install of PDL 2.4.7 on Linux. My > wife has a large spare partition on her desktop. If we make a list of > "target" Linux distributions, I could install each of those from > scratch and see what it takes to get PDL 2.4.7 running. So the install > would be tested on a "pristine" system. > > What distributions would you include? It can't be a huge list. I'm > thinking maybe three: > > - Ubuntu (maybe a separate test for Debian). > - CentOS (covers Fedora and RHEL). > - OpenSUSE. > > The goal is to pick the distros that are most likely to be used by > potential users. In fact... we might even drop OpenSUSE in favour of > OpenSolaris. I suspect that Solaris is more common in math departments > than SUSE. > > Opinions? > Daniel. _______________________________________________ Perldl mailing list [email protected] http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
