Thanks a lot Mike !!! for detailed look at nightly issue. Tried another build machine like you suggested and it worked :) ...
~ Pradhap.D Mike.Sullivan at sun.com wrote: > >From sfwnv-discuss-bounces at opensolaris.org Wed May 21 08:29:56 2008 > > Since I seem to have found your workspace: > >>>> /export4/pd155743/librsync/proto/root_sparc/usr/apache2/2.2/build/libtool >>>> >>>>> /usr/bin/uname >>>>> SunOS rankine 5.11 snv_87 sun4u sparc SUNW,Sun-Fire >>>>> > > I poked a bit. First, though I haven't looked at your changes I see > this: > > mkdir /export4/pd155743/librsync/proto/root_sparc/usr/bin > mkdir: Failed to make directory > "/export4/pd155743/librsync/proto/root_sparc/usr > /bin"; File exists > *** Error code 2 (ignored) > mkdir /export4/pd155743/librsync/proto/root_sparc/usr/bin > mkdir: Failed to make directory > "/export4/pd155743/librsync/proto/root_sparc/usr > /bin"; File exists > *** Error code 2 (ignored) > > which is not your build problem but it almost looks like you're trying to > make $ROOT/usr/bin instead of letting Targetdirs do it. It could > be you're running librsync's 'make install' and it's doing that, > but please make sure you aren't doing that yourself. > > 2nd note this: > > > >>>>> Things that may affect the build. >>>>> /opt/sfw/bin >>>>> /opt/sfw/lib >>>>> /opt/sfw/include >>>>> > > it could be something in /opt/sfw is breaking you. > It's certainly the case that apache's configure finds > gsed there as usual: > > checking for a sed that does not truncate output... /opt/sfw/bin/gsed > > but I didn't think that caused this kind of problem. > > It also looks like the compile line for apache2/re_operators.c > has -I/usr/include before -I$(ROOT)/usr/include so it _could_ be > that rankine has an older libxml2 on it, though it seems to be > dated April 1 so it's not that old. As far as I can tell that > compile line matches the gates. > > your esp-gs failure seems to be because it doesn't link against X, > but rankine does have recent X bits (should be about the same > as on the gate). And certainly in its configure I see: > > checking for IceConnectionNumber in -lICE... no > checking for XOpenDisplay in -lX11... no > checking for XdbeQueryExtension in -lXext... no > checking for XtAppCreateShell in -lXt... no > > which are all 'yes's in the gate. I don't know if that means something > in your environment caused that or if the X upgrade on rankine > failed and I missed it, and I can't check because there seems to > have been a clobber in your workspace so there's no config.log to > look at. > > you certainly do have extra environment variables too. > > At this point, since you said you tried norm's env -i suggestion, > I'd move to another machine. rankine is an ON build machine and so > has only built ON ever, has been bfu'd, and mostly only bfu'd, > for a long long time. While I started updating bunches of non-ON bits > when it transferred to me (and I started watching SFW), it's entirely > possible that the various upgrade scripts don't account for things like > leaping over multiple releases at a time :) > > Mike > >
