On Wed, Nov 05, 2003 at 08:18:54AM +0100, Rafael Garcia-Suarez wrote: > Michael G Schwern wrote: > > On Wed, Nov 05, 2003 at 12:56:35AM +0100, Rafael Garcia-Suarez wrote: > > > > Undertested commits leads to major breakage. > > > > SDBM_File won't build anymore... > > > > ext/SDBM_File/sdbm/Makefile.PL overrides top_targets but doesn't call the > > superclass. That's really fishy and fragile. In particular, it means > > the blibdirs target never gets written. > > > > Frankly, the whole sdbm directory is fishy. But that's old news. > > Yes indeed. But how many fishy Makefile.PLs are there out there ? > Usually *you*'re a member of the Backward Compatibility Police...
The sdbm/Makefile.PL violated encapsulation by trying to completely reimplement a major MakeMaker method. I'm sympathetic to modules authors breaking encapsualtion to get stuff done because there's often no other way to do it, but sdbm goes too far by not even attempting to give lip-service to SUPER::top_targets(). And without good reason, as far as I can tell. Basically all my patch does is cause SUPER::top_targets() to be called. But, I will grep CPAN to see if anyone's fooling with top_targets. -- Michael G Schwern [EMAIL PROTECTED] http://www.pobox.com/~schwern/ 1. You are thin. 2. You look great. 3. You live in a castle with robot servants. -- http://www.angryflower.com/anaddi.gif
