net-snmp works just fine on systems w/out pci, and all the code in net-snmp is conditional upon pci being available, so hard depending on the package is wrong
If you have a patch that'll help — otherwise it'll have to wait till tomorrow and I'll make it USE-conditional.
i didn't write a patch for configure, but it should be easy to utilize the autoconf cache variables ala (untested): export ac_cv_header_pci_pci_h=$(usex pci) export netsnmp_cv_func_pci_lookup_name_LMIBLIBS=$(usex pci)
Yes I meant a patch to the ebuild. Will do that tomorrow, no problem.
Fixed in 5.7.2_rc1-r1 (since some arches already marked -r0 stable).
thanks, latest one seems to work nicely
*** Bug 436078 has been marked as a duplicate of this bug. ***
I don't know if this bug should be reopened or a new one filed or nothing at all, but please note that with the recent removal of net-analyzer/net-snmp-5.7.2_rc1-r1 the stable version once again is 5.7.2_rc1 thus bringing back the problem on PCI-less systems, after an emerge -u world the snmpd service isn't able to start.
Doesn't have to be reopened. -r1 was never stable to begin with, if you used package.accept_keywords just fix it on your side.
*** Bug 439534 has been marked as a duplicate of this bug. ***
This doesn't appear to be fixed. I've recently synced, /usr/portage/net-analyzer/net-snmp/net-snmp-5.7.2_rc1.ebuild Is shown as stable: KEYWORDS="alpha amd64 arm hppa ia64 ~mips ppc ppc64 s390 sh sparc x86" But it still forces usage of pciutils COMMON="ssl? ( >=dev-libs/openssl-0.9.6d ) tcpd? ( >=sys-apps/tcp-wrappers-7.6 ) rpm? ( app-arch/rpm dev-libs/popt ) bzip2? ( app-arch/bzip2 ) zlib? ( >=sys-libs/zlib-1.1.4 ) elf? ( dev-libs/elfutils ) python? ( dev-python/setuptools ) sys-apps/pciutils lm_sensors? ( sys-apps/lm_sensors )"
(In reply to comment #10) > This doesn't appear to be fixed. > > I've recently synced, > /usr/portage/net-analyzer/net-snmp/net-snmp-5.7.2_rc1.ebuild net-snmp-5.7.2 (release, not release candidate) is in tree since Oct 20. Does this reproduced with it?
If using just 5.7.2 (which is currently unstable) the PCI use flag is respected. Building with USE="-pci" succeeds and snmpd behaves as expected. If security issues are fixed I'd really like it tagged stable.
(In reply to comment #12) > If security issues are fixed I'd really like it tagged stable. Unfortunately, this version can not be stabilized yet and all security issues in our bugzilla are related to versions <5.7.2_rc1, so fast stabilization is not an option. Maybe optional USE="pci" can be applied to _rc1?
Nobody's going to touch the stable ebuild. For the few people needing the pciless version of this, just use ~arch version until this is marked stable (which is going to happen next month anyawy).