The new pysnmp-4.1.6a is keyworded for ~ia64 and ~sparc - it was added to portage today. It should RDEPEND on libsmi, but libsmi currently isn't keyworded for ~ia64 and ~sparc.
Here is a way to test to see that libsmi is working correctly. smilint is a binary that installs with libsmi, and IP-MIB is a standard IETF MIB file that ships with libsmi as well. Run the following command: smilint -s -l 6 /usr/share/mibs/ietf/IP-MIB It should have the following output: /usr/share/mibs/ietf/IP-MIB:4672: [5] warning: use Integer32 instead of INTEGER in SMIv2 /usr/share/mibs/ietf/IP-MIB:4693: [5] warning: use Integer32 instead of INTEGER in SMIv2 /usr/share/mibs/ietf/IP-MIB:4707: [5] warning: use Integer32 instead of INTEGER in SMIv2 /usr/share/mibs/ietf/IP-MIB:4759: [5] warning: use Integer32 instead of INTEGER in SMIv2 /usr/share/mibs/ietf/IP-MIB:2375: [5] warning: index of row `ipAddressPrefixEntry' can exceed OID size limit by 141 subidentifier(s) /usr/share/mibs/ietf/IP-MIB:2587: [5] warning: index of row `ipAddressEntry' can exceed OID size limit by 139 subidentifier(s) /usr/share/mibs/ietf/IP-MIB:2773: [5] warning: index of row `ipNetToPhysicalEntry' can exceed OID size limit by 140 subidentifier(s) /usr/share/mibs/ietf/IP-MIB:3140: [5] warning: index of row `ipDefaultRouterEntry' can exceed OID size limit by 140 subidentifier(s) /usr/share/mibs/ietf/IP-MIB:4641: [5] warning: index element `ipAdEntAddr' of row `ipAddrEntry' should be not-accessible in SMIv2 MIB /usr/share/mibs/ietf/IP-MIB:4740: [5] warning: index element `ipNetToMediaIfIndex' of row `ipNetToMediaEntry' should be not-accessible in SMIv2 MIB /usr/share/mibs/ietf/IP-MIB:4740: [5] warning: index element `ipNetToMediaNetAddress' of row `ipNetToMediaEntry' should be not-accessible in SMIv2 MIB
Created attachment 105158 [details] Failed tests on sparc There are two failing tests for sparc with libsmi-0.4.5, output attached.
Gustavo, Thanks for checking on this. Is there any way you could post the results in the subdirectories that failed? The results are different than on x86, but x86 is the only arch I have access to. I fear an upstream bug here. Thanks.
gustavo: please post those diffs from the test failure. ia64: please test and add your ~keyword.
http://dev.gentoo.org/~gustavoz/sparc/libsmi-0.4.5-sparc-tests.tar.bz2
Uh, the tests are so weird. In ia64 it fails in: FAIL: smidump-cm.test But in x86 it fails in: FAIL: smidump-orig-smiv2.test Interesting how i get the same errors as Gustavo but two different errors in two different machines.
News?
Hi 'smidump-cm.test' is non-critical, it generates dia graphs of the SMI files, and the rightmost corner of a graphical rectangle appears to differ by 1 pixel on some systems. ~arch is safe for that now, i'm just looking for a reasonable way to turn off specifically that test in the build system. Could you please post your failure of smidump-orig-smiv2.test? There should be diff files that it directed you to.
comparing `smidump -f smiv2 SNMPv2-MIB' output with dumps/orig-smiv2/*. comparing `smidump -f smiv2 IF-MIB' output with dumps/orig-smiv2/*. comparing `smidump -f smiv2 MAU-MIB' output with dumps/orig-smiv2/*. comparing `smidump -f smiv2 RMON2-MIB' output with dumps/orig-smiv2/*. *** smidump output differs, see smidump-orig-smiv2.out/*.diff FAIL: smidump-orig-smiv2.test That's what i get. All the test dump files are in my d.g.o tarball.
Ok, the tests are just overly pedantic. The order of a couple of items in teh while changed. Sorry to take so long to get back to you. You can just keyword for those failures you gave me, they are really fine.
~ia64 done, then.
~sparc'd and tests restricted on 0.4.5.