I just tried to emerge bmon-2.1.0-r2 (new emerge) and found out that the libnl update (from 0.5.0->1.0_pre6) broke this package since the headers below aren't on the system anymore. The error starts out like this, and there's much more past this. in_netlink.c:47:29: error: netlink/helpers.h: No such file or directory in_netlink.c:51:34: error: netlink/route/filter.h: No such file or directory in_netlink.c:55: error: variable 'nl_h' has initializer but incomplete type in_netlink.c:56: error: variable 'link_cache' has initializer but incomplete type
Indeed... That's why it was masked before. See bug 140207 for more information. Anyway, that means now we have to fix bmon ;-).
libnl-1.0_pre6 still broken with bmon-2.1.0-r2 as bmon still need some missing C headers from libnl-0.5.0 Shouldn't libnl-1.0_pre6 still masked as this problem is not solved yet and even bmon is not patched to look at the right *new files*? Add packages to /etc/portage/package.mask is not the way things should be. Right now I'm masking libnl onto some servers as it's easier than make a patch for bmon.
*** Bug 191124 has been marked as a duplicate of this bug. ***
*** Bug 191126 has been marked as a duplicate of this bug. ***
Here's a workaround... http://www.linuxquestions.org/questions//showthread.php?t=494468 if test $ac_cv_lib_nl_nl_connect = yes; then If someone knows how to modifie this if statement so it will check for the libnl-1.0 version, so it can be disabled. Or maybe just check in the ebuild for >= libnl-1.0 and than just patch away the if statement... so bmon compiles without libnl... (configure Line 7328) cat >>confdefs.h <<_ACEOF #define HAVE_NL "1" _ACEOF LIBNL="-lnl" NL="Yes" else case ${target_os} in *linux*) echo echo "*** Warning: Building bmon on Linuxx w/o libnl ***" echo ;; esac fi
Sry I missed some line: if test $ac_cv_lib_nl_nl_connect = yes; then cat >>confdefs.h <<_ACEOF #define HAVE_NL "1" _ACEOF LIBNL="-lnl" NL="Yes" else case ${target_os} in *linux*) echo echo "*** Warning: Building bmon on Linuxx w/o libnl ***" echo ;; esac fi
Created attachment 129942 [details, diff] bmon-2.1.0-r2.ebuild.patch Okay, here's a very very very _very_ dirty and quick, _very quick_ patch for the current ebuild
Created attachment 129944 [details, diff] bmon-2.1.0.patch Okay, here's a very very very _very_ dirty and quick, _very quick_ patch for bmon
Created attachment 129947 [details, diff] bmon-2.1.0-libnl-1.0.patch Sry, wrong patch attached...
builds sanely for me with this patch applied as -r3 though. please test
Hmm... well, it works, yeah, but uhm... I'd rather set it to UNRESOLVED and wait for a *real* fix (i.e. a working libnl implementation) My two cent -Ionic
net-analyzer/bmon-2.1.0-r3 USE="-dbi -rrdtool" seems to be fine on both amd64 and x86 while the current stable version does not even build out of the box (see comment 0).
Stable on x86
seems to work fine on amd64, stabled
This is done long ago