On Sun, 03 Oct 2004 23:19:57 -0400 Michael wrote:
Mike,

Thanks for taking the initiative on this. Comments below. I agree with your
assessment on the other items.


MJS> 1.Remove SO_BSDCOMPAT socket option setting from
MJS>     snmplib/snmpUDPDomain.c and from
MJS>     snmplib/snmpUDPIPv6Domain.c
MJS> 
MJS>     These were useful only in Linux 2.0 kernels ?
MJS>     Seems low risk and appropriate.  Comments ?

If it is kernel specific, then the kernel version needs to be tested for, and
the code should be ifdef'd out, not removed.

MJS> 8.Comment out "No MemShared line in /proc/meminfo" in
MJS>    agent/mibgroup/ucd-snmp/memory.c
MJS> 
MJS>    Which bug does this fix?

I think RedHat has this same patch. It's a log message that should probably be
a debug message. Or even better, track down which kernel versions have the
memshared line, and ifdef the expectation of this line. But it's much easier to
simple convert to a debug msg.

MJS> 9.Various LmSensors changes [NOT DONE]
MJS>    agent/mibgroup/ucd-snmp/lmSensors.c
MJS>       MIB object changes (Gauge to Integer) violate RFC 1902
MJS>    mibs/LM-SENSOR-MIB.txt
MJS>       MIB object changes (Gauge to Integer) violate RFC 1902

I should have dug a little deeper... RFC 1902 has been obsoleted by RFC 2578
(STD 58). The prohibition against changing object definitions did not change.


MJS> 11.Patch to ignore interfaces that have no statistics
MJS>    agent/mibgroup/mibII/interfaces.c
MJS> 
MJS>    Is there documentation change needed?
MJS>     Seems low risk and appropriate.  Comments ?

Hmmm... How do we know the difference between an interface that wont
ever have statistics, and one that doesn't currently have statistics? I don't
think the MIB specifies any such requirement for inclusion in the table.

I'd suggest that excluding interfaces should be user-configured, not part of
the base code. ie, a conf token to exclude interfaces seems more appropriate.

MJS> 12.Patch to log level Error -> Debug when printing data format error
MJS>    agent/mibgroup/mibII/interfaces.c
MJS> 
MJS>    This doesn't seem right. Comments ?

I agree. The format should be well known, and if there is a parse error, it
should be logged. If the format has changed, the code should be updated to
handle the new version.

I can see, however, perhaps adding a static variable to only log the message
once per run, so that a malformed entry doesn't log a message every time the
table is hit.

MJS> 17.Perl command addition to substitute /usr/local/etc/snmp for 
MJS> SNMPCONFPATH in
MJS>     local/Makefile.in
MJS> 
MJS>     Seems low risk and appropriate.  Comments ?

Shouldn't that already be handled by configure? We shouldn't rely on perl being
installed.


MJS> 18.Include /etc/snmp in def_search_difs
MJS>     local/mib2c
MJS> 
MJS>     With change (keep other line referencing /usr/local/etc/snmp),
MJS>     Seems low risk and appropriate.  Comments ?

I think the number of search directories should be kept at a minimum. The
default should be the location where the conf files are installed, which is the
current behaviour.  In 5.2, it is possible to set the environment variable
MIB2C_DIRS to include additional paths.


MJS> 21.Add HAVE_DLOPEN
MJS>     acconfig.h
MJS>     configure.in
MJS> 
MJS>     Seems low risk and appropriate.  Comments ?

I'd like to know what this is fixing..

MJS> 22.FAQ changes /usr/local --> /usr

Hmmm.. the net-snmp default install prefix is /usr/local.


MJS> 23.Seven new MIBs and one existing MIB are installed
MJS>    mibs/LM-SENSORS-MIB.txt  (existing)

I believe the LM-SENSORS-MIB is only installed if lmSensors support is
configured in.

MJS>    mibs/Makefile.in

what does this makefile do?

MJS>    Other MIBs (except LM-SENSORS-MIB)
MJS>    *should* be obtainable from IETF MIB repository. Comments?

I'm not sure what criteria Wes has for adding MIBs to the distribution...


MJS> 25.configure.in
MJS>     --without-rpm   Don't use RPM even if available.

ok.

MJS>     Add -lcrypto if EVP_md5 is available in crypto library
MJS>     Add -ldl  if dlopen function is defined
MJS>     Modify HAVE_SSIZE_T to provide value "1"

hmmm... I thought configure already dealt with these. What is being changed?



MJS> 28. Add "-- core" to clean rule in Makefile.rules
MJS> 
MJS>      Why ?

It's normal to remove core files, but not sure what the '--' is for.


-- 
Robert Story; NET-SNMP Junkie <http://www.net-snmp.org/>
<irc://irc.freenode.net/#net-snmp>
Archive: <http://sourceforge.net/mailarchive/forum.php?forum=net-snmp-coders>

You are lost in a twisty maze of little standards, all different. 


-------------------------------------------------------
This SF.net email is sponsored by: IT Product Guide on ITManagersJournal
Use IT products in your business? Tell us what you think of them. Give us
Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out more
http://productguide.itmanagersjournal.com/guidepromo.tmpl
_______________________________________________
Net-snmp-coders mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/net-snmp-coders

Reply via email to