rbd@lamont.Columbia.edu (roger davis) (03/23/88)
I'm having a very fundamental (and probably stupid) problem bringing up R2 on a Sun 3/180 running 3.4. The installation crashes very early on. Following is the entire output from 'make World': cd ./util/imake; make clean; make rm -f ccflags imake rm -f *.CKP *.ln *.BAK *.bak *.o core errs ,* *~ *.a tags TAGS make.log \#* cc -o ccflags ccflags.c cc -o imake imake.c `./ccflags` mv Makefile Makefile.bak make -f Makefile.bak Makefile rm -f Makefile.bak; mv Makefile Makefile.bak mv: Makefile: Cannot access: No such file or directory *** Error code 1 (ignored) ./util/imake/imake -TImake.tmpl -I./util/imake.includes -s Makefile -DTOPDIR=. make Makefiles Making Makefiles in ./include... rm -f ./include/Makefile.bak mv ./include/Makefile ./include/Makefile.bak mv: ./include/Makefile: Cannot access: No such file or directory *** Error code 1 (ignored) cd ./include; .././util/imake/imake -TImake.tmpl -I.././util/imake.includes -s Makefile -DTOPDIR=./..; \ make -b Makefiles sh: ./include: bad directory *** Error code 1 Stop. *** Error code 1 Stop. *** Error code 1 Stop. I've done only minimal editing of Sun.macros (just setting DESTDIR to an absolute pathname), but tried the original version of the file as well, to no avail. I've seen several postings from other Sun sites saying that installation went without a hitch, so I'm assuming that I did something wrong. I picked the distribution up via ftp the day that our site got the release announcement -- was there a problem with imake or any makefiles that was fixed just after the announcement? The top-level Makefile looks suspicious to me: World:: cd $(IMAKESRC); $(MAKE) clean; $(MAKE) -$(MV) Makefile Makefile.bak $(MAKE) -f Makefile.bak Makefile $(MAKE) Makefiles $(MAKE) clean $(MAKE) depend $(MAKE) $(WORLDOPTS) ... Makefile:: $(IMAKE) Makefile:: Imakefile \ $(IRULESRC)/Imake.tmpl \ $(IRULESRC)/Imake.rules \ $(IRULESRC)/$(MACROFILE) -$(RM) Makefile.bak; $(MV) Makefile Makefile.bak $(IMAKE_CMD) -DTOPDIR=$(TOP) When the 'make -f Makefile.bak Makefile' happens, the '$(RM) Makefile.bak; $(MV) Makefile Makefile.bak' fails because '-$(MV) Makefile Makefile.bak' has already been done. It goes downhill fast from that point. This was great fun to debug, considering that the imake that's done immediately after the failed mv not only destroys the original Makefile, thus covering its tracks, but also replaces it with an even worse Makefile which contains no define for $(MV). Thus, any subsequent attempt comes back with cd ./util/imake; make clean; make rm -f ccflags imake rm -f *.CKP *.ln *.BAK *.bak *.o core errs ,* *~ *.a tags TAGS make.log \#* cc -o ccflags ccflags.c cc -o imake imake.c `./ccflags` Makefile Makefile.bak Make: Cannot load Makefile. Stop. Make: Cannot open Makefile.bak. Stop. and really leaves you scratching your head. What have I done wrong?