On Tue, Feb 19, 2013 at 7:10 PM, Benjamin Kaduk = <ka= duk@mit.edu> wrote:

We've had a few commits over the years to help parallel builds along by= splitting rules with two targets into two separate rules, one a pure depen= dency and the other with the actual contents of the rule (usually compile_e= t).

A typical example is 147aeeb6c59b5f4a9f8a795a91e0c42ecf80278c (gerri= t/257) (trimmed):
-budb_errs.c budb_client.h: budb_errs.et budb_client.p.h
+budb_client.h: budb_errs.c
+
+budb_errs.c: budb_errs.e= t budb_client.p.h
=A0 =A0 =A0 =A0 $(RM) -f budb_client.h budb_errs.c; ${COMPILE_ET} -p ${srcd= ir} budb_errs -h budb_client

However, these changes seem to cause problems for FreeBSD make, even for se= rial builds. =A0The build log shows a failure to install budb_client.h , No= such file or directory, but then continues on to compile linktest before b= ailing out (sorry for the poorly wrapped copy/paste):

install =A0-o root -g wheel -m 444 JUAFS/libjuafs.a /usr/ports/net/openafs/= work/openafs-1.6.2/lib/libjuafs.a
install =A0-o root -g wheel -m 444 UAFS/libuafs.a /usr/ports/net/openafs/wo= rk/openafs-1.6.2/lib/libuafs.a
case amd64_fbsd_91 in alpha_dux*|sgi_*|sun4x_*|sunx86_*|rs_aix*|*lin= ux*|hp_ux11*|ia64_hpux*|*_darwin_1*|*nbsd*|*obsd*|*fbsd*) cd = src && cd shlibafsauthent && make all ;; =A0*) =A0echo Not = building shared libafsauthent for amd64_fbsd_91 ;; =A0esac
install =A0-o root -g wheel -m 444 budb_client.h /usr/ports/net/openafs/wor= k/openafs-1.6.2/include/afs/budb_client.h
install: budb_client.h: No such file or directory
*** [/usr/ports/net/openafs/work/openafs-1.6.2/include/afs/bu= db_client.h] Error code 71
cc -g -O -I/usr/ports/net/openafs/work/openafs-1.6.2/src/config -I/u= sr/ports/net/openafs/work/openafs-1.6.2/include -I. -I. =A0-O2 -pipe= -fPIC =A0 -c budb_errs.c
[ yes !=3D "" ] || case amd64_fbsd_91 in alpha_dux*|sgi_*|sun*_5*= |rs_aix*|*linux*|hp_ux11*|ia64_hpux*|*[nof]bsd*) cd src &= & cd tptserver && make all ;; =A0*_darwin_[1-6][0-9]) =A0echo N= ot building MT ptserver for amd64_fbsd_91 ;; =A0*_darwin_*) =A0cd src &= & cd tptserver =A0&& make all ;; =A0*) =A0echo Not building MT = ptserver for amd64_fbsd_91 ;; =A0esac

The make debugging output is not immediately enlightening, but provides som= e hints.

budb_server (???) depends on ${TOP_INCDIR}/afs/budb_client.h depends= on budb_client.h depends on budb_errs.c depends on budb_errs.et. Yet budb_errs.c has been gener= ated from the rule for budb_errs.h (${COMPILE_ET} -p ${srcdir} budb_errs) a= nd is up-to-date, so make's internal bookkeeping sees the empty rule to= make budb_client.h from budb_errs.c, claims succes (I don't see this i= n the debugging log, though?!) and tries to install the header.

The problem seems to stem from there being two calls to compile_et using budb_errs.et, one of wh= ich makes budb_errs.h and the other which makes budb_client.h. =A0Swapping = the order in which the multiple targets are split out, so that budb_errs.c = depends on budb_client.h which is made via compile_et, lets my build finish= here, but it's probably not the best solution.

Is there reason to not make a budb_client.h that is the contents of budb_cl= ient.p.h with #include <budb_errs.h> added? =A0Alternately we could p= rocess the template manually with the shell.

[tbudb has the same issue]

-Ben