[Bioperl-l] Bioperl 1.5.2 RC5 install onWinXPActivePerl 220.127.116.119
cjfields at uiuc.edu
Fri Dec 1 11:49:32 EST 2006
On Dec 1, 2006, at 3:26 AM, Nathan S. Haigh wrote:
> In addition, using CPAN allows you to run the test suite easily
> without the need to download it separately and run it after a PPM
A PPM, by design, is supposed to imply that the distribution passes
tests for the specified platform, at that point in time, after all
prereqs are installed and any additional postinstall operations
(install C libraries, modify config files, etc) are complete. The
ActiveState automated PPM building process dictates that; if it fails
any test, it will not be made into a PPM. It's sort of a 'stamp of
approval' that all tests pass, so you don't need to run them.
However, a test may fail (and a PPM may not get generated) for pretty
superficial reasons, such as the makefile not specifying that a
module is needed, server issues, etc, so the automated process isn't
fullproof. That's why Kobes and the other repositories are
available, where the PPM/PPD is manually generated and made to work
specifically for Windows (or whatever other platform).
Saying that, it is completely up to the person packaging the
distribution to follow those rules if one were to make a PPM
manually. You don't even have to run tests prior to using 'nmake
ppd'. We can currently state, though, that all tests pass when all
prereqs are installed for this distribution. At least at this point
> I don't know of a way to clean out ActivePerl - I use VMWare
> Workstation and have a virtual machine with a fresh install of
> WinXP and ActivePerl 18.104.22.1689 - maybe someone else has ideas?
I haven't tried it that way. I have Parallels on Mac OS X (I run a
SigmaPlot/Excel combo off it). My tests were using a native WinXP
installation (i.e. not virtually) on my old Dell. It shouldn't make
a difference; VMWare, Parallels, and the like should all run
ActivePerl for WinXP since it's a virtual machine. Windows Vista, on
the other hand...
I think with PPM4 you can install to a custom directory. It may be
possible to install all new modules to that directory, then you would
at least have an idea of what was there (though I don't think you can
delete it directly w/o screwing up the PPM database).
More information about the Bioperl-l