On Saturday 01 December 2007 06:47:23 James Keenan via RT wrote: > I wove what I was doing on this RT into what I was doing in > http://rt.perl.org/rt3/Ticket/Display.html?id=47902 and have submitted > the results in a patch attached to that ticket: > http://rt.perl.org/rt3/Ticket/Attachment/324326/144446/diff.noConfig.patch. >txt > > I did not implement a command-line option for 'osname' because (a) I > don't think we need it quite yet and (b) when we do implement it, it > will probably be in a configuration step of its own. > > In other respects, I tried to follow the intent of Aldo's original > submission. Once committed, we will take a reading on $^O in the second > configuration step, init::defaults, place it in a part of the > Parrot::Configure object where we store other readings from the Perl > environment, and then look up that value at points in later > configuration steps where we have taken a direct reading on $^O.
We'll probably have to perform a similar replacement in the remaining tests, at least to make cross-compilation test correctly. This is a good step toward that. -- c
