graphics/bmeps fails to configure, if databases/gdbm is installed. configure picks up gdbm, but cannot use it: checking gdbm.h usability... yes checking gdbm.h presence... no configure: WARNING: gdbm.h: accepted by the compiler, rejected by the preprocessor! configure: WARNING: gdbm.h: proceeding with the compiler's result checking for gdbm.h... yes Since configure has found gdbm, it will skip the check for ndbm. Later, it finds out that it cannot use either gdbm or ndbm and fails: ERROR: None of the supported database APIs was found on this system. At least one of NDBM, GDBM or Berkeley DB is required. Fix: The patch attached creates a patch that has configure skip the gdbm check and use ndbm whether or not gdbm is present. (Note that graphics/bmeps/files has to be created.) Patch attached with submission follows: How-To-Repeat: Install databases/gdbm and try to build graphics/bmeps.
Maintainer of graphics/bmeps, Please note that PR ports/164029 has just been submitted. If it contains a patch for an upgrade, an enhancement or a bug fix you agree on, reply to this email stating that you approve the patch and a committer will take care of it. The full text of the PR can be found at: http://www.freebsd.org/cgi/query-pr.cgi?pr=ports/164029 -- Edwin Groothuis via the GNATS Auto Assign Tool edwin@FreeBSD.org
State Changed From-To: open->feedback Awaiting maintainers feedback (via the GNATS Auto Assign Tool)
The upstream site says: In May 2010 I merged several formerly separated projects into the dktools project. I hope this results in easier maintenance for me and in easier installation for the users to have the formerly ``dklibs'', ``bmeps'', ``fig2vect'', ``jlayout'', ``genau'', ``dktools'', ``fchksum'', ``tcptool'', ``hostsadm'', ``fsnmp'', ``yanolc'' and ``prqd'' projects in one package now. As you can see, the problem is that the distfile has a bunch of totally unrelated programs all together. So the configure script is trying to find libraries for all of them. I get this message: ERROR: None of the supported database APIs was found on this system. At least one of NDBM, GDBM or Berkeley DB is required. NDBM is shipped as part of the OS for some UNIX systems. GDBM is the Gnu Database Manager, this software can be found on GNU software mirrors, i.e. ftp://ftp.informatik.rwth-aachen.de/pub/gnu/gnudbm/ . The Berkeley DB is an open source embedded database system. SleepyCat is now owned by Oracle, I found the sources at http://www.oracle.com/database/berkeley-db/ but the URL may change if Oracle reorganizes the web site. WARNING: The netsnmp library is missing. This library is recommended. Some programs (i.e. fsnmp, klpinfo and snmpyalc) can not be built without this library. Use your favorite package management method to install the library _and_ the developer support. If there is no package available for your system you can attempt to install from source. See http://sourceforge.net/projects/netsnmp/ for the sources. But, the build doesn't actually fail because we only build bmeps, which shouldn't need databases or snmp. Can you ignore the message and continue, or does it completely fail? -- Rob Farmer
On 01/13/2012 00:07, Rob Farmer wrote: > But, the build doesn't actually fail because we only build bmeps, > which shouldn't need databases or snmp. Can you ignore the message and > continue, or does it completely fail? What do you mean with "the build doesn't actually fail"? cd /usr/ports/graphics/bmeps/ make fails (on 9.0-RELEASE/amd64 with current ports tree and databases/gdbm installed) with the error message you got, too, and additionally this: ===> Script "configure" failed unexpectedly. Please report the problem to rfarmer@predatorlabs.net [maintainer] and attach the "/usr/ports/graphics/bmeps/work/dktools-2.2.22/config.log" including the output of the failure of your make command. Also, it might be a good idea to provide an overview of all packages installed on your system (e.g. an `ls /var/db/pkg`). *** Error code 1 Stop in /usr/ports/graphics/bmeps. I do not know how to ignore this. Tools dealing with ports consider this a build failure. Yes, I can got to work/dktools-2.2.22/ and run make successfully, but this is not how ports are supposed to be used. Since you say that bmeps does not actually use databases, my patch will do not harm, but fix the build (of the port). Have I misunderstood you? Cheers, Jan Henrik
On Thu, Jan 12, 2012 at 3:35 PM, Jan Henrik Sylvester <me@janh.de> wrote: > On 01/13/2012 00:07, Rob Farmer wrote: >> >> But, the build doesn't actually fail because we only build bmeps, >> which shouldn't need databases or snmp. Can you ignore the message and >> continue, or does it completely fail? > > > What do you mean with "the build doesn't actually fail"? > It prints a message saying "ERROR," but the configure script finishes, exits with 0 status, and the build works: ERROR: None of the supported database APIs was found on this system. At least one of NDBM, GDBM or Berkeley DB is required. NDBM is shipped as part of the OS for some UNIX systems. GDBM is the Gnu Database Manager, this software can be found on GNU software mirrors, i.e. ftp://ftp.informatik.rwth-aachen.de/pub/gnu/gnudbm/ . The Berkeley DB is an open source embedded database system. SleepyCat is now owned by Oracle, I found the sources at http://www.oracle.com/database/berkeley-db/ but the URL may change if Oracle reorganizes the web site. WARNING: The netsnmp library is missing. This library is recommended. Some programs (i.e. fsnmp, klpinfo and snmpyalc) can not be built without this library. Use your favorite package management method to install the library _and_ the developer support. If there is no package available for your system you can attempt to install from source. See http://sourceforge.net/projects/netsnmp/ for the sources. If the configure script complains about a missing library but you are sure the library (including developer support) is installed, make sure to have the CFLAGS and LDFLAGS variables set up properly. On Unix/Linux systems you may want to use CFLAGS="-I/usr/local/include" LDFLAGS="-L/usr/local/lib" export CFLAGS export LDFLAGS On Mac OS you may want to use CFLAGS="-I/sw/include" LDFLAGS="-L/sw/lib" export CFLAGS export LDFLAGS if you have Fink installed. ===> Building for bmeps-2.2.22 >>> 2012-01-13 02:22:55 Compile b2main.c cc -DHAVE_CONFIG_H -I. -I. -I/usr/local/include -O2 -fno-strict-aliasing -pipe -march=nocona -o b2main.o -c b2main.c In file included from ./dktypes.h:125, from ./dk.h:96, from bmepsi.h:53, from b2main.c:66: /usr/include/sys/termios.h:3:2: warning: #warning "this file includes <sys/termios.h> which is deprecated, use <termios.h> instead" > cd /usr/ports/graphics/bmeps/ > make > > fails (on 9.0-RELEASE/amd64 with current ports tree and databases/gdbm > installed) with the error message you got, too, and additionally this: > > ===> Script "configure" failed unexpectedly. > Please report the problem to rfarmer@predatorlabs.net [maintainer] and > attach > the "/usr/ports/graphics/bmeps/work/dktools-2.2.22/config.log" including the > output of the failure of your make command. Also, it might be a good idea to > provide an overview of all packages installed on your system (e.g. an `ls > /var/db/pkg`). > *** Error code 1 > > Stop in /usr/ports/graphics/bmeps. > > I do not know how to ignore this. Tools dealing with ports consider this a > build failure. > > Yes, I can got to work/dktools-2.2.22/ and run make successfully, but this > is not how ports are supposed to be used. > > Since you say that bmeps does not actually use databases, my patch will do > not harm, but fix the build (of the port). > > Have I misunderstood you? > I can't reproduce the problem with just gdbm installed. I can reproduce it with net-snmp so I have expanded the patch to cover that. Removing the gdbm part didn't work for me, but I left it since it won't hurt anything. -- Rob Farmer
I just wondered why this bug is still in the feedback state. I thought it was awaiting maintainer feedback, which it got 1.5 month ago. Or is it waiting for another submitter feedback as the maintainer suggested a larger patch? For the record: The port still fails for me without any patch. Rob Farmer's patch works for me. (In my setting, it produces exactly the same package as with my patch.) Rob Farmer's patch should be committed and as he is the maintainer, no more feedback should be needed for that. Thanks, Jan Henrik
State Changed From-To: feedback->open new patch received as feedback.
Responsible Changed From-To: freebsd-ports-bugs->scheidell I'll take it.
scheidell 2012-03-03 13:28:05 UTC FreeBSD ports repository Modified files: graphics/bmeps Makefile Added files: graphics/bmeps/files patch-configure Log: - Unbreak with net-snmp - Bump PORTREVISION PR: ports/164029 Submitted by: Jan Henrik Sylvester <me@janh.de> Reviewed by: Rob Farmer <rfarmer@predatorlabs.net> (maintainer) Approved by: gabor (mentor, implicit) Revision Changes Path 1.16 +1 -0 ports/graphics/bmeps/Makefile 1.1 +108 -0 ports/graphics/bmeps/files/patch-configure (new) _______________________________________________ cvs-all@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/cvs-all To unsubscribe, send any mail to "cvs-all-unsubscribe@freebsd.org"
State Changed From-To: open->closed Committed. Thanks!
scheidell 2012-03-05 18:27:56 UTC FreeBSD ports repository Modified files: graphics/bmeps/files patch-configure Log: - Fix previous patch mangled by committer (scheidell, me) PR: ports/164029 Submitted by: Jan Henrik Sylvester <me@janh.de> Approved by: Rob Farmer <rfarmer@predatorlabs.net> (maintainer, implicit), gabor (mentor, implicit) Revision Changes Path 1.2 +101 -108 ports/graphics/bmeps/files/patch-configure _______________________________________________ cvs-all@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/cvs-all To unsubscribe, send any mail to "cvs-all-unsubscribe@freebsd.org"