[When I posted this report yesterday, I failed to cc perl6-internals, and it doesn't seem to
 be 
propagating to that list.  So I'm re-posting it; see rt for the attachments.  Sorry for any 
duplication.]

Once again I am unable to successfully 'make' Parrot on Darwin.  I  
getting the same errors I was getting when I first tried at the  
Chicago hackathon on Nov 11 -- nothwithstanding (a) having succeeded  
with 'make' since then several times; and (b) applying all the hack  
workarounds suggested at the hackathon.

I did 'make realclean', then took two approaches depending on how I  
called 'perl Configure.pl'.  In both approaches, I then applied  
chromatic's hack to Makefile:  deleting '-bundle' from the values  
assigned to LD_LOAD_FLAGS (approx line 164).  But 'make' failed in  
both approaches, though it got a few lines farther in the second.

perl Configure.pl
# no options
# See output in 09.12.2006.make.failure.output.txt	

perl Configure.pl --without-gmp --cc=gcc --ccflags=-DAN
# options suggested as workarounds at hackathon
# See output in 09.12.2006.make.failure.output.2.txt
	

# See differences in two outputs of failed 'make' in diff. 
09.12.2006.make.failure.output.txt

I must say I'm getting very frustrated with this inability to get  
Parrot to 'make' successfully.  I'm using the same OS and hardware  
(Darwin, iBook G4) as many other Parrot hackers.  I've upgraded to  
the most recent version of GNU make (3.81).  I do a 'make realclean'  
and an 'svn update' every time before running Configure.pl.  And  
still, a month into this, I cannot get 'make' to succeed successfully.

I'm hoping to contribute to the project by phalanxing its build  
tools, pmc2c.pl in particular.  The ultimate test of any refactoring  
I do on pmc2c.pl is:  Does 'make', which calls pmc2c.pl repeatedly,  
succeed at least as well when running with my refactored version as  
with the trunk version?  But how can I make that determination when I  
cannot get 'make' to work even with the trunk version?

kid51