On Thu, May 30, 2013 at 06:37:51PM +0300, Konstantin Belousov wrote: > On Thu, May 30, 2013 at 10:29:33AM -0500, Brooks Davis wrote: > > On Thu, May 30, 2013 at 10:53:23AM -0400, John Baldwin wrote: > > > On Thursday, May 30, 2013 10:09:58 am Brooks Davis wrote: > > > > Author: brooks > > > > Date: Thu May 30 14:09:58 2013 > > > > New Revision: 251140 > > > > URL: http://svnweb.freebsd.org/changeset/base/251140 > > > > > > > > Log: > > > > Always define INSTALL_DDIR and define it such that it contains no > > > > extra > > > > / characters rather than removing them later on. This should fix > > > > release builds. > > > > > > > > PR: conf/178963 > > > > Reviewed by: gjb, hrs > > > > > > > > Modified: > > > > head/Makefile.inc1 > > > > > > > > Modified: head/Makefile.inc1 > > > > ============================================================================== > > > > --- head/Makefile.inc1 Thu May 30 14:07:14 2013 (r251139) > > > > +++ head/Makefile.inc1 Thu May 30 14:09:58 2013 (r251140) > > > > @@ -522,7 +523,7 @@ _includes: > > > > @echo > > > > "--------------------------------------------------------------" > > > > @echo ">>> stage 4.1: building includes" > > > > @echo > > > > "--------------------------------------------------------------" > > > > - ${_+_}cd ${.CURDIR}; ${WMAKE} SHARED=symlinks par-includes > > > > + ${_+_}cd ${.CURDIR}; ${WMAKE} par-includes > > > > _libraries: > > > > @echo > > > > @echo > > > > "--------------------------------------------------------------" > > > > > > Did you mean to change this? It seems unrelated. > > > > Oops, I forgot that I had that in for testing. Reverted in 251141. > > > > That said, the SHARED=(copies|symlinks) code should probably go on the > > chopping block. It's a lousy idea except for populating WORLDTMP and > > here it saves at most 1.5% in /usr/obj. Worse, the implementation is > > hand coded in each Makefile so it adds complexity for a mostly useless > > feature. > > I disagree with the statement about the useless, at least for WORLDTMP > it saves the full buildworld when a change in the usermode headers is > done during the development. E.g., changing include/* or sys/sys/* > still allows to build libc with buildenv, instead of spending 40+ > minutes on the buildworld.
Ok, that's fair. I won't touch the includes bits. I seriously doubt 40+ minutes given that -DNO_CLEAN should work fine in such a case. Do you agree that it's less than useful in the never exercised cases in share/examples and share/sendmail? -- Brooks
pgppNwsMH1b66.pgp
Description: PGP signature