The release_cached_resources() routine only release caches that are registered
with an OID. Is this done for a reason? Since when one tries to create a
cache_handler without specifying the root oid of the cache, this new cache is
not linked to the global cache_handler: cache_head.
Many thank
On Mon, 25 Jun 2007 00:29:23 +0200
Thomas Anders <[EMAIL PROTECTED]> wrote:
> Dave Shield wrote:
> > Does such a setup define "SIZEOF_LONG_LONG" ?
> > If so where? (And *why*?!)
>
> So what happens to the Windows builds if you remove "#define
> SIZEOF_LONG_LONG 8" from win32/net-snmp/net-snmp-co
> "TA" == Thomas Anders <[EMAIL PROTECTED]> writes:
TA> I've reviewed them and they look safe:
Your definition of safe and mine are likely different...
I'd say 1.5.23b is probably safer because it's been out a while (feb).
But 1.5.24 isn't even listed on their website and is only 9 days old.
Wes Hardaker wrote:
>> "TA" == Thomas Anders <[EMAIL PROTECTED]> writes:
>
> TA> I've just updated SVN trunk to libtool 1.5.24 [1] (from 1.5.22).
> TA> Is it worth considering to do the same for 5.4.x before 5.4.1.rc1?
>
> Have you reviewed the changelog/deltas/announcement/whatever?
I've
> "TB" == Thijs Brouwer <[EMAIL PROTECTED]> writes:
TB> Ok, looks like it would be logical to change all three of them when
TB> any renaming is done. But having such generic defines in a header
TB> seems dangerous. At one point or another, there's going to be
TB> trouble...
Yeah, we've slowly
> "AC" == Ajay Chenampara <[EMAIL PROTECTED]> writes:
AC> the problem was with snmpUtils.pm that came with cricket - it was
AC> passing a "default" context. got rid of it and all is good -
AC> Thanks!!
Ugh. That's a horrible choice as a default.
--
Wes Hardaker
Sparta, Inc.
--
ML> My apologies ahead of time for the vintage of net-snmp that I've
ML> applied this to. The patch is relatively small, and should hopefully
ML> be easily adaptable to more recent versions.
We appreciate the submission none the less (we'd love a more recent
patch, but we will take what we can ge
> "MF" == Magnus Fromreide <[EMAIL PROTECTED]> writes:
MF> The question that rise from this is how to handle sysORTable in
MF> subagents, the obvious way seems to be that sysORTable entries are
MF> automatically reregistered just like mib registrations are.
Probably the best choice... cache
> "AB" == Alex Burger <[EMAIL PROTECTED]> writes:
AB> Patch 1728247 added to types.h:
Better question... where are these used? I don't believe there is any
place in the code that a long long should be used at this point. IE, if
it's in there it's likely because broken code was put in somew
> "DS" == Dave Shield <[EMAIL PROTECTED]> writes:
DS> We'll probably need to issue a 5.2.4.1 release to fix this problem
DS> on that line too.
Wanna do that with just a delta of a patch to this issue? If so, I'd
say just do it with at most a single rc1 pre-release. Or none at all...
--
Wes
> "TA" == Thomas Anders <[EMAIL PROTECTED]> writes:
TA> I've just updated SVN trunk to libtool 1.5.24 [1] (from 1.5.22).
TA> Is it worth considering to do the same for 5.4.x before 5.4.1.rc1?
Have you reviewed the changelog/deltas/announcement/whatever? I'd be
hesitant to replace it so late
> "BS" == Bruce Shaw <[EMAIL PROTECTED]> writes:
BS> One of the Open Solaris mailing lists (sfwnv-discuss) just had a flame
BS> war about libtool. It appears its use is deprecated for a number of
BS> reasons I didn't necessarily follow.
Cause it is a pain in the neck to deal with, that's wh
Thomas Elsgaard wrote:
> From 5.4 net-snmp is now default requiring perl, BUT that has been
> forgotten in the description in the INSTALL file for 5.4 and 5.4.1, i
> hope that this is already corrected in the next versions, but haven't
> checked..
It doesn't really *require* Perl (since if you don
One of the Open Solaris mailing lists (sfwnv-discuss) just had a flame
war about libtool. It appears its use is deprecated for a number of
reasons I didn't necessarily follow.
Given that it appears we're stuck with it for the moment (I believe
autoconf needs it) I'd want to have the most up-do-d
Hello
From 5.4 net-snmp is now default requiring perl, BUT that has been
forgotten in the description in the INSTALL file for 5.4 and 5.4.1, i
hope that this is already corrected in the next versions, but haven't
checked..
See the release notes for 5.4:
Important Changes:
- The defa
I am at the point where I am integrating the NetSNMP package into our system's
build environment. I perform the "./configure" with the approriate compiler
and linker options. After this completes, I perform a successful "make" as
well. However, when the "make install" is performed I get a pe
Hi,
First, is there any maximum packet size in any snmp version (1, 2c or 3)?
Furthermore, is there a maximum packet size when using agentx? Because I am
building an agentx sub-snmp and when I try to send a large packet nothing is
received at the manager station but when I had reduced the pack
On 26/06/07, ALBERT BEZZINA <[EMAIL PROTECTED]> wrote:
> With the above setting, I don't think that the cache_free function is called
> automatically
Checking the code, it looks as if the routine release_cached_resources()
should be called every 60s, to release any expired caches. You'd have to
e
Hi,
I am writing an snmp agent, and I am trying to use the cache_handler.
I have a question:
I have the following conditions:
no flags set and the 'timeout' value is assigned to 10.
With the above setting, I don't think that the cache_free function is called
automatically, so is it possible to
On Tue, 26 Jun 2007 10:09:36 +0300 Juuso wrote:
JA> On Monday 25 June 2007 19:22:45 Juuso Alasuutari wrote:
JA> > Next up in my small monitor app project is to learn how to get snmp data
JA> > from devices at constant intervals. [...]
JA> > init_snmp( "foo" )
JA> >
JA> > while {
JA> > for each de
Ugh, yet another self-solving problem... After spending 5 minutes looking at
snmpget.c I noticed that I hadn't reset the len variable between read_objid()
calls. Sorry for the noise (again)! At least writing all these "bug reports"
helps focus my attention.
Juuso
On Tuesday 26 June 2007 13:11:
I noticed something strange when reading two OIDs with a small client program
[0]. The values are returned properly only when the OIDs are added to the PDU
in a specific order.
If I request HOST-RESOURCES-MIB::hrSystemProcesses.0 (type Gauge32) before
DISMAN-EVENT-MIB::sysUpTimeInstance (type T
On Monday 25 June 2007 19:22:45 Juuso Alasuutari wrote:
> Next up in my small monitor app project is to learn how to get snmp data
> from devices at constant intervals. I expanded the example application
> (http://www.net-snmp.org/wiki/index.php/TUT:Simple_Application) into
> something that can be
23 matches
Mail list logo
