net-analyzer/sussen-0.90 doesn't compile with RPM 5.0

[Jeff Johnson]

tracker

[Alexander Skwar (a.skwar)]

Tried to rebuild sussen with these flags:

[ebuild N ] net-analyzer/sussen-0.90 USE="-debug doc gnome" 0 kB

Fails:

[...]
 i686-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I. -I../../.. -I../../.. -march=pentium-m -O2 -pipe -fomit-frame-pointer -MT interface.lo -MD -MP -MF .deps/interface.Tpo -c interface.c -fPIC -DPIC -o .libs/interface.o
if /bin/sh ../../../libtool --mode=compile i686-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I. -I../../.. -I../../.. -march=pentium-m -O2 -pipe -fomit-frame-pointer -MT rpm.lo -MD -MP -MF ".deps/rpm.Tpo" -c -o rpm.lo rpm.c; \
        then mv -f ".deps/rpm.Tpo" ".deps/rpm.Plo"; else rm -f ".deps/rpm.Tpo"; exit 1; fi
 i686-pc-linux-gnu-gcc -DHAVE_CONFIG_H -I. -I. -I../../.. -I../../.. -march=pentium-m -O2 -pipe -fomit-frame-pointer -MT rpm.lo -MD -MP -MF .deps/rpm.Tpo -c rpm.c -fPIC -DPIC -o .libs/rpm.o
In file included from rpm.c:3:
/usr/include/rpm/rpmlib.h:11:20: error: rpmtag.h: No such file or directory
/usr/include/rpm/rpmlib.h:12:24: error: rpmversion.h: No such file or directory
In file included from rpm.c:3:
/usr/include/rpm/rpmlib.h:122: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'rpmHeaderFormats'
/usr/include/rpm/rpmlib.h:264: error: expected ')' before 'first'
/usr/include/rpm/rpmlib.h:291: error: expected declaration specifiers or '...' before 'Header'
[...]

I'll attach the complete build log.

Is that because of some incompatibility with RPM 5.0? I have that installed with the following flags:

[ebuild R ] app-arch/rpm-5.0.0 USE="berkdb -bzip2 -doc -file -lua -neon nls pcre -perl python (-selinux) -sqlite" 0 kB

--($:~)-- emerge --info
Portage 2.1.4_rc14 (default-linux/x86/2007.0/desktop, gcc-4.2.2, glibc-2.7-r1, 2.6.23-gentoo-r1.r08.mit-ide-mod_2 i686)
=================================================================
System uname: 2.6.23-gentoo-r1.r08.mit-ide-mod_2 i686 Genuine Intel(R) CPU T2400 @ 1.83GHz
Timestamp of tree: Tue, 08 Jan 2008 08:46:01 +0000
distcc 2.18.3 i686-pc-linux-gnu (protocols 1 and 2) (default port 3632) [disabled]
ccache version 2.4 [enabled]
app-shells/bash: 3.2_p33
dev-java/java-config: 1.3.7, 2.1.3
dev-lang/python: 2.4.4-r4, 2.5.1-r5
dev-python/pycrypto: 2.0.1-r6
dev-util/ccache: 2.4-r7
sys-apps/baselayout: 1.12.10-r5
sys-apps/sandbox: 1.2.18.1-r2
sys-devel/autoconf: 2.13, 2.61-r1
sys-devel/automake: 1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r2, 1.10
sys-devel/binutils: 2.18-r1
sys-devel/gcc-config: 1.4.0-r4
sys-devel/libtool: 1.5.24
virtual/os-headers: 2.6.23-r3
ACCEPT_KEYWORDS="x86 ~x86"
CBUILD="i686-pc-linux-gnu"
CFLAGS="-march=pentium-m -O2 -pipe -fomit-frame-pointer"
CHOST="i686-pc-linux-gnu"
CONFIG_PROTECT="/etc /usr/kde/3.5/env /usr/kde/3.5/share/config /usr/kde/3.5/shutdown /usr/share/config"
CONFIG_PROTECT_MASK="/etc/env.d /etc/env.d/java/ /etc/fonts/fonts.conf /etc/gconf /etc/php/apache2-php5/ext-active/ /etc/php/cgi-php5/ext-active/ /etc/php/cli-php5/ext-active/ /etc/revdep-rebuild /etc/terminfo /etc/udev/rules.d"
CXXFLAGS="-march=pentium-m -O2 -pipe -fomit-frame-pointer"
DISTDIR="/Gentoo/Portage/distfiles"
EMERGE_DEFAULT_OPTS="--alphabetical"
FEATURES="buildpkg ccache collision-protect distlocks metadata...

[Alexander Skwar (a.skwar)]

Created attachment 140660
net-analyzer\:sussen-0.90\:20080110-200836.log

Build log.

[Alexander Skwar (a.skwar)]

Seems like sussen 0.90 is incompatible with RPM 5.0. I downgraded to rpm-4.4.6-r4, and now I'm able to build sussen-0.90.

[pva (pva)]

Mike, may be I'm completely misunderstood, but seems that headers exported by rpm-5.0 are not usable as some of them depend on other system headers they do not include and some are just incomplete. At least I did not manage to include anything but rpm/rpmio.h, while sussen needs rpmps.h included too. If I am correct, then may it's worth to hardmask rpm-5.0? Take a look at bug 202399 too.

[Ma3oxuct (ma3oxuct)]

rpm-5 is not backwards compatible with older rpm versions, which would explain the cause of this bug as well as the other bug Peter cited. Hard masking until sussen and net-snmp are "made up to speed" with rpm-5 certainly won't hurt.

[Alexander Skwar (a.skwar)]

Just curious, but wouldn't it be better to just change sussen and net-snmp to block when RPM 5 is installed?

[Mike Frysinger (vapier)]

no, adding blockers just makes peoples' lives miserable

i dont see how this could have worked against rpm 4 considering it didnt actually install any headers ... which means sussen has some automagic detection running that the ebuild is not correctly controlling ?

[pva (pva)]

(In reply to comment #6)
> i dont see how this could have worked against rpm 4 considering it didnt
> actually install any headers ...

As far as I see, it does. In any case, rpm-5.0 is new package which breaks compatibility with the previous releases. Both net-snmp and sussen require changes to work with new version, may be, as described here:

http://<email address hidden>/msg00113.html

Until this changes occur (I'm more interested in net-snmp fixes ATM) the best we can do is hardmask rpm-5.0.

Mike, any objections for hardmasking? I think in gentoo world net-snmp is more important package then rpm ;)

[Mike Frysinger (vapier)]

unless you can explain to me how it works against rpm-4.x without any actual rpm headers installed, i'm not OK with p.masking

i'm not going to argue the relative merits of rpm vs net-snmp as you're most likely correct, but masking one package so that you can ignore bugs in another doesnt sound like the route to go for any set of packages

[pva (pva)]

Created attachment 141992
rpm-4.4.6-r3 installed file list

(In reply to comment #8)
> unless you can explain to me how it works against rpm-4.x without any actual
> rpm headers installed, i'm not OK with p.masking

Mike, rpm-4.4.6-r3 does install headers. Take a look at attachment. This is file list on my system...

> i'm not going to argue the relative merits of rpm vs net-snmp as you're most
> likely correct, but masking one package so that you can ignore bugs in another
> doesnt sound like the route to go for any set of packages

Sure. But issue was reported upstream and rpm developers are aware about breakage and are going to help us fixing this issues. Until this happens it's better to avoid breakage of unstable tree.

[Mike Frysinger (vapier)]

please consult latest rpm-4.4.7 ... istr that was the version i checked out and it wasnt installing any headers ... but that sounds like a different issue we need to address ... which means masking rpm-5.x wont fix breakage as rpm-4.4.7 is no good

if you check that out, we can mask 5.0.0 and open another bug for 4.4.7 ...

[pva (pva)]

(In reply to comment #10)
> please consult latest rpm-4.4.7 ... istr that was the version i checked out and
> it wasnt installing any headers ...

Ah, now I see. Only rpm-4.4.7-r3 which is currently package masked does not install headers. All previous revisions/versions do. Also on x86 ~rpm-4.4.7 is masked too at default-linux/x86/package.mask. So I've masked >=app-arch/rpm-4.4.7-r3 for now. I hope I've manage resolve this issues soon...

[Jeff Johnson (n3npq)]

There's a patch to the oval/unix/rpm.c file that will permit compiling with rpm-5.x here:

    http://wraptastic.org/pub/jbj/sussen-rpm.c-rpm5.patch

You will also need to add -I/usr/include/rpm/ to CFLAGS in the Makefile.

Yes the patch is quite raw. Much better needs to be done, but that will require participation
from sussen developers.

[Jeff Johnson (n3npq)]

tracked at https://bugs.launchpad.net/rpm/+bug/634179

[pva (pva)]

Hi Jeff. Could you attach patch here? The link you posted here is not available any more...

[pva (pva)]

*** Bug 370951 has been marked as a duplicate of this bug. ***

[Wfdawson (wfdawson)]

I just bumped into this issue from revdep-rebuild after installing rpm. Any hope of seeing it resolved, or should I just punt and zap sussen?

[Jeff Johnson (n3npq)]

The disk on which the patch resided crashed years ago.

Lemme see if I can find ... meanwhile sussen wasn't in good shape
in 2008, and there are many changes to rpm's API since, all versions.

[Pacho-gentoo (pacho-gentoo)]

dropped