bogus autoconf tests and missing pieces result in libnetsnmp.so not being linked
to libm, causing breakage that is most noticable in collectd/cups-libs ports
when linking with ld.bfd. bump major due to removed functions and add comments
to Makefile with more info.
Enable installing snmpps/snmptop, they have been around for a while but were
just segfaulting in tests before - turns out this was actually only when
pointed at OpenBSD snmpd which doesn't handle hrSWRunPerfMem/hrSWRunPerfCPU
and was OK pointed at net-snmp snmpd - so patch to fix the segfault.
- enable pre-RFC AES192/256 support ("--enable-blumenthal-aes"),
diff from martijn@ (adds symbols so I added a library minor bump for
this).
- disable hidden dep on lpstat from cups; if present this resulted in
adding a function to libnetsnmpmibs so a major bump is needed for this.
- recognize openbsd 7.x
there may be some missing as my unpacked ports source is a little out of date
but this should catch the main things people might run into
the struct was reordered a second time in sysctl.h r1.192 to improve
compatibility but amd64 snapshot packages made it out before that happened
so the bumps are still needed
too far off, and I'm fed up with this mega-diff sitting in my tree gathering
C's :)
add a flavour "readonly" which uses --enable-read-only, this disables all code
handling SETs.
a local IP address is set. Cause is a use-after-free, OpenBSD's malloc
helpfully often picks these up early (with relatively low impact on
performance), often turning what would be a hidden bug into a likely
crash.
Reported and analysis from Andrea Biscuola of Zabbix,
https://support.zabbix.com/browse/ZBX-12726https://sourceforge.net/p/net-snmp/bugs/2803/
Diff from me. I was going to wait for confirmation from Andrea that it
fixes the issue, but having read more of the code I'm confident that this
change or something like it is needed, so committing now to make it pre 6.2.
ok naddy
Theodore Wynnychenko. This tells it not to exit when it can't open
/dev/kmem etc. Fixes startup now that this access is disabled by
default in OpenBSD.
Some MIBs won't work, but in general these should be rewritten
to use sysctl interfaces.
user has set a command-line arguments containing '+' (not totally uncommon).
Otherwise rc_check() now waits for pexp to match, which fails due to
the + being handled as part of a regular expression.
disabling the ipv6 mib - this is done in the rc script so it can be
overridden with netsnmpd_flags if required (but just run snmpd from base,
really ....)
Reported by Pierre Kim.
Drop unnecessary patch while there, the problem it was meant to fix was
already fixed in 5.7.3.
updates are triggered. This is following the struct if_data ABI change a few
days ago; if_msghdr has an embedded if_data. Some may be unnecessary, but
some are definitely needed and bumps are cheaper than debugging.
Problem reported with wpa_supplicant by Mikolaj Kucharski.
while traversing inpt_queue. Fix botch in previous commit by
initializing next and prev once each, rather than prev twice and next
not at all. Eliminate 'skip:' and use 'continue' instead, since code
after 'skip:' was pointless. Nuke now unused variable 'head'.
