Date: Mon, 12 Apr 2004 17:27:39 +0200 From: Herve Quiroz <herve.quiroz@esil.univ-mrs.fr> To: Greg Lewis <glewis@eyesbeyond.com> Cc: Kris Kennaway <kris@obsecurity.org> Subject: Re: [ports-i386@bento.FreeBSD.org: spread-j-3.17.1 failed on i386 4] Message-ID: <20040412152739.GA14027@arabica.esil.univ-mrs.fr> In-Reply-To: <20040409225932.GC8348@misty.eyesbeyond.com> References: <20040409004930.GT42056@xor.obsecurity.org> <20040409225932.GC8348@misty.eyesbeyond.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Apr 09, 2004 at 04:59:32PM -0600, Greg Lewis wrote:
> On Thu, Apr 08, 2004 at 05:49:30PM -0700, Kris Kennaway wrote:
> > compile:
> > [javac] Compiling 12 source files to /tmp/a/ports/net/spread-j/work/spread-src-3.17.1/java/dest/class
> >
> > BUILD FAILED
> > /tmp/a/ports/net/spread-j/work/spread-src-3.17.1/java/build.xml:44: Error running jikes compiler
>
> Hmmm, the port is trying to use jikes but has no dependency on it. Did it
> ever build? I think the port should set USE_JIKES.
You're right:
[from build.xml:]
<property name="build.compiler" value="jikes"/>
so the port should enforce USE_JIKES=yes
(BTW, I tought Ant would only use jikes if it were available...)
> However, it looks like this one also exposes a bug in bsd.java.mk. I
> can't see that USE_JIKES sets up a BUILD_DEPENDS for jikes. Herve?
Okay, I found the problem. However, actually dealing with the
JAVA_{BUILD|RUN} and NO_{BUILD|RUN}_DEPENDS_JAVA is quite a PITA, as
their behaviors are quite "contradictory". IMHO, the best method would
be to remove completely support for NO_{BUILD|RUN}_DEPENDS_JAVA. Still,
there are quite a lot of ports that use the old scheme:
$ cd /usr/ports
$ find . -name Makefile -mindepth 3 -maxdepth 3 -exec grep -l "_DEPENDS_JAVA" {} \;
./databases/arcexplorer/Makefile
./dns/crossip/Makefile
./graphics/aoi/Makefile
./java/junit/Makefile
./java/xdoclet/Makefile
./net/edonkey-gui-java/Makefile
./net/freenet/Makefile
./net/frost/Makefile
./net/java-beepcore/Makefile
./net/phex/Makefile
./net/xnap/Makefile
./textproc/fop/Makefile
./textproc/dtd2relax/Makefile
./textproc/dtdinst/Makefile
./textproc/jing/Makefile
./textproc/ppower4/Makefile
./textproc/saxon/Makefile
./textproc/xt/Makefile
./textproc/trang/Makefile
I can find something if you want. There's another way though: patch all
of these ports to use the new scheme ...but then there are 19 ports to
fix, which will take some time.
Hum. Maybe a quick vi, ed, or sed script would do the trick. I'll check
that.
Herve
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040412152739.GA14027>