On Friday 16 November 2007 17:51:52 James Keenan via RT wrote: > Apart from the *general* problem of relying too heavily on Perl 5'c > Config.pm (via init::defaults) for values during testing of the > configuration step tests, does the following accurately summarize the > reasons why *specific* tests are failing on your systems?
> chromatic: > Is using a vendor-supplied Perl that has libgdbm setting within its > Config.pm and is using that Perl to run Parrot configuration. However, > libgdbm doesn't exist on his system. Hence, failures at point(s) where > it is assumed that that library is present. > > Andy Dougherty: > Is using a vendor-supplied Perl has 'cc' in its Config.pm. However, he > also has gcc present on his system and is using that C-compiler to build > Parrot. Hence, failures at point(s) where it is assumed that 'cc' is > the C-compiler being used to build Parrot. This is accurate for my system and looks accurate for Andy's system as well. -- c
