On Mon Mar 05 16:57:47 2007, [EMAIL PROTECTED] wrote: > Two months ago, I filed this bug report (excerpt):
I'm still not seeing the effects of the bug you describe, btw. (on OS X intel or ppc). It does remind me of issues I had when attempting to use the GMP libs, though. > > ...I inferred that some change in the configuration process in the past 3 > > days had caused Configure.pl to override my environmental setting of > > 'ld' to '/usr/bin/g++-3.3'. > > > > Running an 'svn diff' on the last two versions of Configure.pl (see > > attached recent.changes.Configure.pl.txt), I saw: > > > > Index: Configure.pl > > ========================================== > > --- Configure.pl (revision 16268) > > +++ Configure.pl (revision 16416) > > @@ -464,9 +464,9 @@ > > init::defaults > > init::install > > init::miniparrot > > + inter::progs > > init::hints > > init::headers > > - inter::progs > > inter::make > > inter::lex > > inter::yacc > > > > ... indicating that step 'inter::progs' is to be performed > > immediately before 'init::hints', rather than two steps afterwards. > > I peered into config/inter/progs.pm (for the very first time) and > > noted that it has a provision for setting the value of 'ld'. config/ > > inter/progs.pm itself had not been changed in the last 3 days, > > therefore I hypothesized that my problems were *solely* to the patch > > applied to Configure.pl on Jan 04. > > > > Once I ran 'make realclean' and 'svn update', I then manually > > reverted Configure.pl to have 'inter::progs' done two steps after > > 'init::hints'. I then examined the resulting 'myconfig' and saw that, > as I desired, line 12 once again ran: > > > > ld='/usr/bin/g++-3.3' > > > > I then ran 'make', which DWIMmed. ... > > > > On the basis of the above, I recommend that Configure.pl be reverted > > to r16268. (If that cannot be done, then I would like to know how to > > call Configure.pl so that I can 'make' Parrot successfully.) > > > > The bug has not been addressed. Up until the last few days, I was > content to let it go. But now I'm worried that it may be the cause of > the bug t/pmc/object-mro.t I reported this morning > (http://rt.perl.org/rt3/Ticket/Display.html?id=41704). Do you still get the bug when you revert the change to Configure.pl? > Thoughts? > > kid51 I tried reverting the change in a local copy of svn head today: seems to work the same for me either way, so unless another platform objects, let's switch it back if it fixes things for you.