Comment 1 Boris Samorodov 2006-08-02 23:15:08 UTC

Responsible Changed
From-To: freebsd-ports-bugs->bsam

Take.

Hi Karel!


I've taken your PR/99993. Great work, thanks!

I have two questions to discuss.

1. The port's Makefile assumes that not only i386 ARCH may be used. So
   far we don't have a non-i386 bootstrapping tools for other ARCHS to
   compile the port. I.e. when I try to compile the port at amd64, the
   ports says that it's only for i386. ;-)

Should we use ONLY_FOR_ARCHS= i386 knob and delete other ARCH stuff?
Otherwise there will be errors while building the port at the ports
building cluster.

2. You use "@unexec /bin/rmdir %D/lib/gcc 2>&1 || true" at the plist
   file. It's OK from the package building script point of view
   (i.e. no error is generated). But from a
   committer/maintainer/developer point of view those messages may be
   misunderstood when looking for (potential) problems with the port.
   BTW, the standard @dirrmtry macro uses "2>/dev/null" instead of
   "2>&1" statement. I don't insist, it's only my opinion. It's up
   to you how to deal with your port. ;-)

Other than that seems to be OK.

You may send the new version of the port directly to me and I'll
commit it after some testing.

Thanks once more.


WBR
-- 
Boris Samorodov (bsam)
Research Engineer, http://www.ipt.ru Telephone & Internet SP
FreeBSD committer, http://www.FreeBSD.org The Power To Serve

> 2. You use "@unexec /bin/rmdir %D/lib/gcc 2>&1 || true" at the plist
>    file.

This should be:
2. You use "@unexec /bin/rmdir %D/lib/gcc 2>&1 || true" at the
   ${WRKSRC}/PLIST.lib file.

Comment 4 Boris Samorodov 2006-08-03 14:37:45 UTC

State Changed
From-To: open->feedback

Awaiting for a feedback.

Here is the proposed patch:
--- Makefile.orig	Thu Aug  3 15:45:21 2006
+++ Makefile	Thu Aug  3 17:41:24 2006
@@ -69,11 +69,7 @@
 				GNU_HOST=${CONFIGURE_TARGET} \
 				SUFFIX=${gcc_suffix}
 
-.if ${ARCH} == "amd64"
-CONFIGURE_TARGET=	x86_64-portbld-freebsd${OSREL}
-.elif ${ARCH} == alpha
-USE_GCC=	3.3+
-.endif
+ONLY_FOR_ARCHS=	i386
 
 # You need a compiler who calls an existing GNAT compiler (3.15 or greater).
 # By default, we'll look for one in the path.  If we can't find one, we'll
@@ -167,10 +163,10 @@
 			${FIND} $${d} -type d | ${SORT} -r | ${SED} -e 's/^/@dirrm /g' >>${WRKDIR}/PLIST.lib ; \
 		fi ; \
 	done
-	${ECHO_CMD} "@unexec ${RMDIR} %D/lib/gcc/${CONFIGURE_TARGET} 2>&1 || true" >> ${WRKDIR}/PLIST.lib
-	${ECHO_CMD} "@unexec ${RMDIR} %D/lib/gcc 2>&1 || true" >> ${WRKDIR}/PLIST.lib
-	${ECHO_CMD} "@unexec ${RMDIR} %D/libexec/gcc/${CONFIGURE_TARGET} 2>&1 || true" >> ${WRKDIR}/PLIST.lib
-	${ECHO_CMD} "@unexec ${RMDIR} %D/libexec/gcc 2>&1 || true" >> ${WRKDIR}/PLIST.lib
+	${ECHO_CMD} "@unexec ${RMDIR} %D/lib/gcc/${CONFIGURE_TARGET} 2>/dev/null || true" >> ${WRKDIR}/PLIST.lib
+	${ECHO_CMD} "@unexec ${RMDIR} %D/lib/gcc 2>/dev/null || true" >> ${WRKDIR}/PLIST.lib
+	${ECHO_CMD} "@unexec ${RMDIR} %D/libexec/gcc/${CONFIGURE_TARGET} 2>/dev/null || true" >> ${WRKDIR}/PLIST.lib
+	${ECHO_CMD} "@unexec ${RMDIR} %D/libexec/gcc 2>/dev/null || true" >> ${WRKDIR}/PLIST.lib
 	cd ${WRKDIR} ; ${SED} -i -e "/PLIST.lib/ r PLIST.lib" ${TMPPLIST}
 
 .include <bsd.port.post.mk>

If you approve it I'll commit the port.

Boris, I completely agree with your reasoning and
approve the changes you have done. Thank you.

Comment 7 Boris Samorodov 2006-08-03 21:36:11 UTC

State Changed
From-To: feedback->closed

Committed, thanks!