Conference smurf::buildhelp

1490.0. "URGENT! Can t build ptos.nightly for six weeks now!" by AOSG::FILTER (Automatic Posting Software - mail to flume::puck) Wed May 31 1995 04:43

Date Of Receipt: 	31-MAY-1995 02:58:40.37
From: 	SMURF::FLUME::"[email protected]" "31-May-1995 0257"
To: 	[email protected]
CC: 	
Subj: 	URGENT! Can't build ptos.nightly for six weeks now!

Hi Build Help,

Both me and another member of our UNX team have been attempting to build a
"kernel sandbox" against ptos.nightly since the BL6 pool was opened some
six ago (without success). I gave up after two weeks of early attempts. (I
[and he] are currently backed aganist BL5, since that is the only "tree"
that will "build".)

For "nightly", I would (attempt) to do the following steps:
	Blow completely away my existing sandbox with:
		mksb -undo sbname

	Make a new one with:
		mksb -back ptos.nightly -dir /sbdir -verbose sbname

	Set up env with:	
		workon -sb sbname

	Populate with:
		mklinks -dironly ./ (while ./ = /sbdir/sbname/src)

	Then attempt to build with:
		cd kernel
		build setup BINARY FLAMINGO GENERIC

Not one has ever succeeded while backed against ptos.nightly. When I
called for help (early on) you guys would say something like I need
to remove some (unknown) "Makefile" in "my" sandbox. (How can I have
any such Makefile(s), if I have just completely blown away my sandbox?)

I just gave up after about two weeks (as well as my coworker in our
group).


I attempted once more (yesterday afternoon) and got the following error (very
early):

makepath np/. && cd np &&  exec make 'RELEASE_OPTIONS=-idfile `genloc /src/setup/osf1_idlist`'    MAKEFILE_PASS=BASIC dopass_all
np: created directory
sed  -e '' </usr/sde/osf1/build/ptos.nightly/src/kernel/io/dec/np/npadapter.h >./npadapter.h.X
mv -f ./npadapter.h.X ./npadapter.h
sed  -e '' </usr/sde/osf1/build/ptos.nightly/src/kernel/io/dec/np/npport.h >./npport.h.X
mv -f ./npport.h.X ./npport.h
sed  -e '' </usr/sde/osf1/build/ptos.nightly/src/kernel/io/dec/np/npscs.h >./npscs.h.X
mv -f ./npscs.h.X ./npscs.h
sed  -e '' </usr/sde/osf1/build/ptos.nightly/src/kernel/io/dec/np/npsysap.h >./npsysap.h.X
mv -f ./npsysap.h.X ./npsysap.h
[ /kernel/include/io/dec/pci ]
makepath pci/. && cd pci &&  exec make 'RELEASE_OPTIONS=-idfile `genloc /src/setup/osf1_idlist`'    MAKEFILE_PASS=BASIC dopass_all
pci: created directory
sed  -e '' </usr/sde/osf1/build/ptos.nightly/src/kernel/io/dec/pci/pci.h >./pci.h.X
mv -f ./pci.h.X ./pci.h
sed  -e '' </usr/sde/osf1/build/ptos.nightly/src/kernel/io/dec/pci/pci_debug.h >./pci_debug.h.X
mv -f ./pci_debug.h.X ./pci_debug.h
sed  -e '' </usr/sde/osf1/build/ptos.nightly/src/kernel/io/dec/pci/pnvram_reg.h >./pnvram_reg.h.X
mv -f ./pnvram_reg.h.X ./pnvram_reg.h
sed  -e '' </usr/sde/osf1/build/ptos.nightly/src/kernel/io/dec/pci/pcia.h >./pcia.h.X
mv -f ./pcia.h.X ./pcia.h
[ /kernel/include/io/dec/scs ]
makepath scs/. && cd scs &&  exec make 'RELEASE_OPTIONS=-idfile `genloc /src/setup/osf1_idlist`'    MAKEFILE_PASS=BASIC dopass_all
scs: created directory
sed  -e '' </usr/sde/osf1/build/ptos.nightly/src/kernel/io/dec/scs/sca.h >./sca.h.X
mv -f ./sca.h.X ./sca.h
sed  -e '' </usr/sde/osf1/build/ptos.nightly/src/kernel/io/dec/scs/scamachmac.h >./scamachmac.h.X
mv -f ./scamachmac.h.X ./scamachmac.h
sed  -e '' </usr/sde/osf1/build/ptos.nightly/src/kernel/io/dec/scs/scaparam.h >./scaparam.h.X
mv -f ./scaparam.h.X ./scaparam.h
Make: Don't know how to make scs.h.  Stop.
*** Exit 1
Stop.

(I believe that this is same problem I reported some six weeks ago - it looks
familar! I do remember it was an include file that started with a "s".)

I assumed that everything would be OK after BL 6 was closed an released.
But this is now becoming a very very serious problem.
Do you want me to "submit" to BL7 without even "comnpiling" (more
less "testing") against that "tree"?

I don't doubt that the problems I (and my coworker) are experiencing
are indeed Makefile problems. That is, it looks like someone "defuncted" (or
added) one or more files and didn't drop an updated Makefile. But isn't it a
problem when the "tree" dosn't have the right Makefiles? I suspect that
people who "backed against nightly sometime ago - and did their "setup" then"
don't have this problem.

Please do as I do, that is, create a new "sandbox", and try and "build" and
resolve these problems.

Floyd,