Here's what I've done.
All the .exists targets are gone.
dir_target() which generated them is gone.
The "die if we can't find perl.h" check is gone.
I've replaced it all with this:
config :: $(FIRST_MAKEFILE) blibdirs
$(NOECHO) $(NOOP)
blibdirs :
$(NOECHO) $(MKPATH) \
$(INST_LIBDIR) $(INST_AUTODIR) \
$(INST_ARCHAUTODIR) $(INST_BIN) \
$(INST_SCRIPT) $(INST_MAN1DIR) \
$(INST_MAN3DIR)
$(NOECHO) $(CHMOD) 755 \
$(INST_LIBDIR) $(INST_AUTODIR) \
$(INST_ARCHAUTODIR) $(INST_BIN) \
$(INST_SCRIPT) $(INST_MAN1DIR) \
$(INST_MAN3DIR)
$(NOECHO) $(TOUCH) blibdirs
So all the blib directories get created in one swoop. I figure why screw
around doing it piecemeal? The blibdirs file is touched so we don't do it
twice (like pm_to_blib).
Everything that had a dependency on a .exists file now has one on blibdirs
instead.
Can anyone think of why/what this would break? Was anyone crazy enough
to depend on dir_target()?
Try the snapshot on makemaker.org, please.
--
Michael G Schwern [EMAIL PROTECTED] http://www.pobox.com/~schwern/
Cottleston, Cottleston, Cottleston Pie.
A fly can't bird, but a bird can fly.
Ask me a riddle and I reply:
"Cottleston, Cottleston, Cottleston Pie."
