Bizarrely, the problem has disappeared.
I ran that same sequence twice, before sending the message, so as to
confirm I was doing exactly what I described.
Afterwards, I started an svn-bisect --min 40200, and got good results
right back up to r40248. I then updated to 251, and got a good result there.
This might mean that some artifact is being created by one of the
earlier revs and left lying around. Or it might mean that my computer
can only compile parrot when the output of C<date> has odd parity.
Anyway, ignore this if you like.
=Austin
Austin Hastings wrote:
I just updated to r40248, hoping to get an update that might fix a
problem I've been seeing.
I did a:
> make clean
> perl Configure.pl
> make.
The make ran through to the point where it fails on ln -sf
blib/lib/libparrot.so.1.4.0 blib/lib/libparrot.so, and so I cp'ed it
as usual. Then I did another
> make
This time, the make didn't run to completion, but instead I got:
make[1]: Entering directory `/mnt/win/Sources/parrot/compilers/tge'
.../../parrot ../../runtime/parrot/library/PGE/Perl6Grammar.pbc
--output=TGE/Parser.pir TGE/Parser.pg
make[1]: *** [TGE/Parser.pir] Segmentation fault
make[1]: Leaving directory `/mnt/win/Sources/parrot/compilers/tge'
It's not at all clear to me why this is happening, and it's recent. I
did an svn update last evening and had no such problems.
FWIW: I'm running Ubuntu/andLinux over Windows XP, and the development
directory is sitting on top of an NTFS partition (which is why ln -s
fails, but doesn't really explain the segfault).
=Austin
_______________________________________________
http://lists.parrot.org/mailman/listinfo/parrot-dev
_______________________________________________
http://lists.parrot.org/mailman/listinfo/parrot-dev
