Hi Bruce,
libtool: link: false cru ...
Are you sure you have /usr/ccs/bin in your path?
See http://comments.gmane.org/gmane.network.net-snmp.devel/23222
+Thomas
--
Android apps run on BlackBerry 10
Introducing the
Niels Baggesen wrote:
The attached patch is a simple fix for this bug, just zeroing the
avail field when it would go negative.
A better fix, after 5.7.2, would be to allow the negative value to
pass through by changing the avail field from unsigned to signed.
The MIB declares the object
Madhu Sudhana Rao wrote:
I upgraded my Ubuntu system to latest 12.04 LTS version, after that I am
getting errors in Net-snmp (below). Looks like some of the libperl.so
binaries got broken. Earlier I was having no issues.
Madhu@desktop-india:~/net-snmp-5.4.2.1/agent$ snmpd -f -Lo
snmpd:
Timo Teräs wrote:
Find attached two important fixes for net-snmp-5.7.1.
* netsnmp-arp-netlink-fix.patch
- fixes a problem with my earlier patch that would not reload
the ARP cache if it was once unloaded
- optimises the reload to happen only if the cache was actually
unloaded
Michael Buchholz wrote:
The new patches will add that binding by interface name and also let the
tools (snmpwalk, snmpget, ...) make use of the new features.
Thanks for the patch! Could you please submit it to
http://www.net-snmp.org/patches so it won't get lost?
+Thomas
Dave Shield wrote:
The hardware-based implementation of the UCD disk table
(ucd-snmp/disk_hw) in v5.6 and above can suffer from
arithmetic overflow when calculating percentage usage
on large disks.
The following simple patch should address this problem:
+1
+Thomas
Tristan Colgate wrote:
I've been doing some work lately with using the mib2c perl support and
Support.pm and have hit a bit of a problem. The getnext table handling
breaks if the index is variable length (e.g. and oid or string index).
The code expects to be able to zero out all the
Dave Shield wrote:
On 27 July 2011 18:25, Vikas Jadhavvikas...@gmail.com wrote:
I am getting below error,
bash-3.00# /usr/sfw/sbin/snmpd -L
dlopen failed: ld.so.1: snmpd: fatal: /usr/sfw/lib/libseaProxy.so: wrong ELF
class: ELFCLASS32
dlopen failed: ld.so.1: snmpd: fatal:
Lomax, Michael P. wrote:
I then did the `make test`. At the bottom of the file you will see that
all but 1 test, the last test, succeeded and then there was an error and
the run “Stop”ped with double error code 2s. This led me to check back
for other errors or warnings during the earlier
Michael P. Lomax wrote:
“configure: WARNING: libperfstat.h: present but cannot be compiled
I've checked in a change that I think should fix this. Please try again
with the current git source (V5-7-patches branch) or wait for 5.7.1.rc1
and report back.
+Thomas
Wes Hardaker wrote:
TA perl/TrapReceiver/const-c.inc
TA and perl/TrapReceiver/const-xs.inc
Those files *aren't* in the net-snmp git repsository.
Maybe because Bart has already removed them? :-)
Building Net-SNMP with Perl enabled always modifies the following git source
files:
perl/TrapReceiver/const-c.inc
and perl/TrapReceiver/const-xs.inc
because they seem to get regenerated during the build. This tends to be a pain
when dealing with the git repo.
Should these files be
Hi Wes,
on 2011-05-25 in r20455 you introduced an empty (bogus) perl/NetSNMP.xs
claiming that this'd help newer Perl to cope with the bundle. However, some
systems (e.g. all AIX systems I've tried) dislike that trick and fail to build
with a linker error like:
[...]
ld -bhalt:4 -bM:SRE
Singh, Vinay wrote:
Has anyone tried to build net-snmp 5.7 on AIX ? It did build for
5.6.1 but getting build error in 5.7. Any help on this is appreciated.
Also, I could never get to compile the code on HP-UX. Has anyone tried
to build net-snmp on HP-UX?
The V5-7-patches branch in
Niels Baggesen wrote:
On Fri, Jul 22, 2011 at 05:00:20PM -0700, Dennis Andrews wrote:
I downloaded net-snmp 5.7 (net-snmp-5.7.tar.gz) and built it with my
normal build environment
What OS?
Based on the post, I'd say RHEL6.
+Thomas
Building trunk or the 5.5/5.6/5.7 branches seems to fail on quite a few
non-Linux systems (AIX, HP-UX and IRIX at least) in
snmplib/transports/snmpUDPBaseDomain.c. Here's the error for HP-UX
11.11/PA-RISC:
cc: /my/src/net-snmp-5.7.dev/snmplib/transports/snmpUDPBaseDomain.c,
line 227: error
Singh, Vinay wrote:
I am using gcc
When I forced a 64-bit compilation, it is giving me error
snmp_client.c:1: sorry, unimplemented: 64-bit mode not compiled in
Has anyone compiled it for 64-bit Solaris x86 ?
The error message means that your gcc copy doesn't support 64-bit compilation.
Singh, Vinay wrote:
Can net-snmp 5.6.1 can be compiled to generate 64-bit binaries for Solaris ?
if yes, what option to pass on “configure “ script. It appears that by
default, only 32-bit binaries are generated even if I am trying to compile on
64-bit machine.
Are you using gcc or Sun cc?
Dave Shield wrote:
The fix (attached) is clearly trivial, and I'd like to include this
in 5.5.1 (although it's not strictly a show-stopper).
+1
Thomas
--
Colocation vs. Managed Hosting
A question and answer guide
Niels Baggesen wrote:
The RHEL4 problem should be fixable by adding two dummy parameters to
the call to pci_lookup_name (the call was made varargs in pciutils 2.2
which is in RHEL5)
That RHEL4 problem is still present in trunk. Does anyone care to commit a fix?
Also, building trunk on HP-UX
On Mon, 03 Jan 2011 09:08:31 +0100, Magnus Fromreide
ma...@lysator.liu.se wrote:
All but RHEL4 looks like they are due to struct sockaddr (from
sys/socket.h) not being defined. This in turn is allowed by posix as
it states that Inclusion of the netinet/in.h header may also make
visible all
Bart Van Assche wrote:
Regarding the above patch: it seems to me that the above patch is
incomplete - the configure.d/... changes are included, but the configure
script changes not.
I'm confident he'll run autoconf after applying the patch, but thanks for
reminding him. :-)
+Thomas
Unfortunately, the multi-platform build status of current SVN trunk leaves
something to be desired. I really hope that none of these errors affect
V5-6-patches.
Comments/fixes appreciated.
AIX 5.3:
making all in /my/build/trunk/aix/snmplib
gmake[1]: Entering directory
Wes Hardaker wrote:
I'd like to include the following changes in the final 5.6.1 release.
Since this (rc2) is expected to be the last release candidate I'd like
to get verification that the patch is a good one even though it's a doc
fix.
+1
+Thomas
Niels Baggesen wrote:
The following patch (for V5-6-patches and trunk) allows to configure
with --disable-debugging
+1
+Thomas
--
Lotusphere 2011
Register now for Lotusphere 2011 and learn how
to connect the dots,
Claus Klein wrote:
Please apply this patch and check the result about my FIXME notes.
Please submit your patch to http://www.net-snmp.org/patches so it won't get
lost. Patches that apply cleanly against SVN trunk usually have a higher chance
to get accepted quickly.
+Thomas
Wes Hardaker wrote:
I'm committing changes to both the autoconf version and the libtool
version (which was required for autoconf 2.68).
Please update http://www.net-snmp.org/wiki/index.php/Build_System#Tool_Versions
when you're done.
+Thomas
Wes Hardaker wrote:
This is a patch to patch 2900078 which fixed and broken sysObjectID.
The patch suggestion comes from Fulko Hew via -coders discussion and
works for me.
Looks sensible. +1.
+Thomas
--
Start
Magnus Fromreide wrote:
The new test framework have generated some new test targets.
This patch adds them to the toplevel makefile as well.
+1 (assumed we're doing rc3 which seems likely)
+Thomas
--
This SF.net
Magnus Fromreide wrote:
The attached diff is from head after regeneration of configure and
net-snmp-config.h.in.
I would like to see it (or an updated version) applied before the next
release is done, is that possible?
+1 (assuming this is just from running the proper toolchain)
+Thomas
Wes Hardaker wrote:
However the patch gets +1 for me
+1 from me as well.
+Thomas
--
This SF.net Dev2Dev email is sponsored by:
Show off your parallel programming skills.
Enter the Intel(R) Threading Challenge 2010.
Wes Hardaker wrote:
However the patch gets +1 for me
+1 from me as well.
+Thomas
--
This SF.net Dev2Dev email is sponsored by:
Show off your parallel programming skills.
Enter the Intel(R) Threading Challenge 2010.
Patch 3058451
(https://sourceforge.net/tracker/?func=detailatid=312694aid=3058451group_id=12694)
says:
Revision 19323 broke SA_REPEAT timers by incorrectly removing all flags instead
of just SA_FIRED.
Wes Hardaker wrote::
Ok, here's a now copy that checks all fopens and fgets:
Error handling gets +2. :-)
What about Robert's proposal to add a DEBUGMSG if the file/process went away?
+Thomas
--
This SF.net Dev2Dev
Wes Hardaker wrote:
(and I tested it to make sure the functionality works as expected)
+1, then.
+Thomas
--
This SF.net Dev2Dev email is sponsored by:
Show off your parallel programming skills.
Enter the Intel(R)
Wes Hardaker wrote:
The obvious patch is:
+1
+Thomas
--
This SF.net Dev2Dev email is sponsored by:
Show off your parallel programming skills.
Enter the Intel(R) Threading Challenge 2010.
Magnus Fromreide wrote:
I'd like to add this pretty obvious patch.
It calls DEBUGMSG before returning instead of after...
The obvious ones are the hardest to find. +1.
+Thomas
--
This SF.net Dev2Dev email is
Wes Hardaker wrote:
The following patch provides proper (D)TLS shutdown and cleanup. It has
been tested using both command line tests as well as the existing TLS
test sub-suite.
The clean up is obviously just nice for memory, but the session
closing on the DTLS side was missed as the
Bart Van Assche wrote:
This patch is not just cleanup because:
- It breaks the MSVC build.
- It changes behavior on the Cygwin and MinGW platforms. On these two
platforms the PATH_SEPARATOR is a colon (':') while Net-SNMP uses the
semicolon (';') as ENV_SEPARATOR for these two platforms. Your
Robert Story wrote:
Sounds reasonable. I'll cache the patch and see what happens with rc2.
To be honest, I think caching has more cons than pros here. rc2 is going to
happen anyway and the later you commit your patch, the later we identify
potential problems. IOW, my vote is to apply your
Wes Hardaker wrote:
To fix an error string which was a cut-n-paste without a replacement in
the new DTLS code.
+1 (what a waste to even do voting on such a trivial change. sometimes policies
feel strange.)
--
Sell
Robert Story wrote:
anyhoo, this patch initializes the temporary variable.
+1
--
Sell apps to millions through the Intel(R) Atom(Tm) Developer Program
Be part of this innovative community and reach millions of netbook
Robert Story wrote:
do you think a CFV is needed for changes to non-default test scripts? i.e.
scripts that won't be run unless explicitly requested by the user?
IMHO no. In case others disagree, consider this a +1 for your patch. :-)
+Thomas
Mygapula, Sudheer IN BLR SISL wrote:
Find the attached capture file for further references.
At a first glance, I'd say the agent on 192.168.2.125 is misbehaving, since it
reports EngineBoots/Time values of 0/0 during engine id discovery and 0/291
when responding to the subsequent query.
HTH,
Dave Shield wrote:
So it doesn't feel too out-of-the-question to tweak these
internal routines to return an indication of whether any of the
named file(s) had been found and processed. I'd argue that
the benefits here would outweigh the cost.
+1
+Thomas
Steve DeLaney wrote:
so I made a patch to agent/mibgroup/agent/extend.c and would like to
submit it for consideration to include in 5.6
Thanks for the patch! Please submit it to http://www.net-snmp.org/patches so it
won't get lost.
+Thomas
Bart,
This patch causes fourteen additional tests to fail on Cygwin.
Sorry for that. trunk or branch? Where exactly do you propose to add #include
net-snmp/agent/agent_module_config.h?
+Thomas
--
This SF.net email
Bart Van Assche wrote:
On Tue, Jun 29, 2010 at 6:28 PM, Thomas Anders
thomas.and...@blue-cable.de mailto:thomas.and...@blue-cable.de wrote:
Where exactly do you propose to add
#include net-snmp/agent/agent_module_config.h?
The safest place to add that include statement IMHO
John Waller wrote:
you'll have to forgive my inexperience with snmp. I was referring to the
'make test' that was failing on every test. I attached the Makefile.
Hopefully you can point me in the right direction. Thanks for all your help
on this.
There's a good chance that you have a
John Waller wrote:
I have downloaded and tried to compile versions 5.5, 5.4.3 and 5.3.3 on
a Solaris 9 SPARC server. None of those will compile. Every single test
fails.
Does anyone have any ideas why?
Not until you care to tell us some details. Within the build directory,
- what's the
John Waller wrote:
libtool: link: ( cd .libs rm -f libnetsnmp.la ln -s
../libnetsnmp.la
: libnetsnmp.la
making all in /usr/local/src/net-snmp/net-snmp-5.5/agent
make: Fatal error in reader: Makefile, line 333: Unexpected end of line seen
Current working directory
Wes Hardaker wrote:
On Tue, 01 Jun 2010 01:27:24 +0200, Thomas Anders
thomas.and...@blue-cable.de said:
TA /my/src/net-snmp-main/testing/fulltests/temptests/T001snmpv1get_simple.t:
No such file or directory at /my/src/net-snmp-5.6svn/testing/RUNFULLTESTS
line 426.
That's an artifact
Magnus Fromreide wrote:
One could obviously imagine an alternate approach where I add options to
configure to set SNMPCONFPATH and SNMPSHAREPATH. Would that solution be
better?
Please go ahead!
+Thomas
--
ThinkGeek
/me wrote:
Thomas Anders wrote:
In the nightly builds make test fails with
fulltests/temptests/T001snmpv1get_olducd.t: No such file or directory at
/my/src/net-snmp-5.6svn/testing/RUNFULLTESTS line 389.
Hopefully fixed in r18756.
This seems to be broken again in current SVN trunk
Bart wrote:
Revision: 18844
http://net-snmp.svn.sourceforge.net/net-snmp/?rev=18844view=rev
Author: bvassche
Date: 2010-05-26 19:46:04 + (Wed, 26 May 2010)
Log Message:
---
CHANGES: Cygwin: libnetsnmpmibs, libnetsnmphelpers and libnetsnmpagent do now
build as
/me wrote:
This change broke the build on AIX and IRIX at least. Their toolchains don't
seem to like building an empty library (libnetsnmphelpers).
FWIW, the most relevant change was actually the revision before that (i.e.
r18843), I think.
+Thomas
Bart Van Assche wrote:
On Mon, May 24, 2010 at 11:47 PM, Thomas Anders
thomas.and...@blue-cable.de mailto:thomas.and...@blue-cable.de wrote:
What's the business case? I suggest to look at this from a users
perspective: what exactly do they get in exchange for a backwards
Dave Shield wrote:
What are the advantages to keeping these two libraries separate?
What (amount of) code out there would only depend on the agent lib (and not the
helpers lib) if the dependencies were clean?
That code (with its dependencies) would benefit from having a smaller
footprint,
Dave Shield wrote:
I'm not sure which (if any) of the extension mechanisms rely on
the helpers library, and which just on the agent lib.
I would certainly be interested to know!
+Thomas
--
Bart Van Assche wrote:
The patch below fixes the circular dependency between libnetsnmphelpers
and libnetsnmpagent. It does this by removing the netsnmp_init_helpers()
call from init_agent() and by adding an explicit call to
netsnmp_init_helpers() after every call to init_agent(). This is a
Omer Zakwrote wrote:
Now I am flabbergasted that anything leading toward such documentation
has been censored from Net-SNMP commits.
Like Wes said: we don't censor useful, properly formatted comments.
+Thomas
--
Magnus Fromreide wrote:
How is the make depend target supposed to work?
If I try it on linux it happily tells me that makedepend is missing, if
I install the X11 package that contains it then it fails to work for me.
If I try it on solaris 10 the it tells me that it needs more than 64k
Dave,
Revision: 18786
http://net-snmp.svn.sourceforge.net/net-snmp/?rev=18786view=rev
Author: dts12
Date: 2010-05-18 14:38:38 + (Tue, 18 May 2010)
Log Message:
---
Working(?) HAL-based version of hrFSTable.
The following compiler warnings are produced by that
Weiwei Zhang wrote:
I have registered a subagent with net-snmp master agent. When building
the master agent, v3 is set. So now all v1/v2c/v3 SNMP message can be
processed by master agent.
*When the agent is running, how to config the master agent to let it
only accept v1 or v2c GET
Bart Van Assche wrote:
Does r18755 build correctly on these platforms ?
Looks like it does. There are quite a few issues with make test, but that's a
different story.
+Thomas
--
Wes Hardaker wrote:
It would be easier to simply wrap the older scripts in a giant I have
no perl so run this simple sh script instead.
Especially since these older scripts have been working without Perl before (and
still do in the branches).
+Thomas
SVN trunk (r18754) throws build errors for all of the non-Linux nightly builds.
I haven't determined what exact commit(s) broke it, but the configure changes
are one likely candidate.
E.g. on Solaris, linking libnetsnmp now misses to specify -lnsl.
On AIX, linking libnetsnmpagent throws a
In the nightly builds make test fails with
fulltests/temptests/T001snmpv1get_olducd.t: No such file or directory at
/my/src/net-snmp-5.6svn/testing/RUNFULLTESTS line 389.
The nightly build systems have two things in common:
- no TAP::Harness
- the build happens outside the source dir
Thomas Anders wrote:
In the nightly builds make test fails with
fulltests/temptests/T001snmpv1get_olducd.t: No such file or directory at
/my/src/net-snmp-5.6svn/testing/RUNFULLTESTS line 389.
Hopefully fixed in r18756.
+Thomas
Bart,
Revision: 18726
http://net-snmp.svn.sourceforge.net/net-snmp/?rev=18726view=rev
Author: bvassche
Date: 2010-05-14 08:42:19 + (Fri, 14 May 2010)
Log Message:
---
libtool bug fix: also pass flags starting with -f to the linker since these
can affect
Robert Story wrote:
I'd like to move the configure prompts for values earlier in the configure
process. Currently they are at the end, which means if you fat-finger a
response (or the configure option that provides the value so that you don't
need a response), you have to re-run configure and
Dave Shield wrote:
I'm therefore calling for votes as to whether this should be applied
to the two lines currently in release freeze.
Looks reasonably safe (and obviously limited to Windows anyway) to me. +1
Also +1 for the earlier two Windows build fixes.
+Thomas
Bart Van Assche wrote:
Since the 5.4 branch is currently in release-candidate freeze mode,
this patch has to be voted upon before it can be applied to the 5.4
branch. Please vote whether or not you want to see this patch applied
on the 5.4 branch.
+1
+Thomas
Dave Shield wrote:
1) Latch disk statistics: no / 54x only / both 54x and 52x
2) Missing privKey crash: no / 54x only / both 54x and 52x
3) VACM best match: no / 54x only / both 54x and 52x
4) MIB dir path logging: no / 54x only / both 54x and 52x
+1 for 1), 2), 3): both 54x and 52x
Magnus Fromreide wrote:
This patch is something I think should be applied overall as it will affect
the default tcp host for master agents on unices as well and make it more
likely that it just works for subagents since they will try to connect to
localhost by default.
Given the insecure
Bart Van Assche wrote:
I'm not sure it's OK to change the V5.4 behavior now since the 5.4
branch is in RC stage (see also
http://net-snmp.sourceforge.net/dev/release-policy.html).
Feel free to call for votes.
+Thomas
Tony Thomas wrote:
What coding style/standard is used in Net-SNMP project?
See the CodingStyle document that ships with the source.
+Thomas
--
Download Intel#174; Parallel Studio Eval
Try the new software tools for
Dave Shield wrote:
I'm attaching a simple patch that walks from .0 until the end of the MIB tree
if given an explicit starting point of '.' or no starting point at all.
Comments?
Even though it's a change in behavior, I must admit I like it.
+Thomas
Mike Walker wrote:
My problem now is just how to set CFLAGS=-xtarget=native
-xarch=native64 for building net-snmp.
If I set it via configure (i.e.: ./configure CFLAGS=-xtarget=native
-xarch=native64), then configure stops with errors relating to c
compiler cannot generate executables.
Mike Walker wrote:
This definitely helps me get further along the path. It correctly
forces Sun Studio 12 and successfully generates a 64bit libnetsnmp.so
file. But make now crashes further in the net-snmp build. Looks like
right after reporting running Mkbootstrap for
Mike Walker wrote:
Has anyone successfully built net-snmp 5.5 for Solaris 10 in 64bit mode?
I've built earlier SVN versions this way without any problems other than that
you need all corresponding 64-bit libraries for your configuration (e.g.
libperl).
As Magnus has said: give it a try and
hashim qaderi wrote:
I keep getting this linking error and not sure what -lz library is.
.
/usr/lib64/gcc/x86_64-suse-linux/4.1.2/../../../../x86_64-suse-linux/bin/ld:
cannot find -lz
On SuSE Linux, this library comes with the zlib rpm package that you may want
to install.
+Thomas
Jan,
Do we have any policy which autoconf/autoheader to use?
http://www.net-snmp.org/wiki/index.php/Build_System#Tool_Versions
+Thomas
--
Download Intel#174; Parallel Studio Eval
Try the new software tools for
Sylvain Dery wrote:
I'm trying to cross-compile the snmp agent for a arm-linux system on a
Fedora 9 system and I get the following error:
/usr/local/arm/ep93xx-d1/lib/gcc/arm-linux/3.4.3/../../../../arm-linux/bin/ld:
ERROR: /tmp/ccm9qwTf.o does not use Maverick instructions, whereas a.out
Bart Van Assche wrote:
Have you considered adding that define in
include/net-snmp/system/solaris.h instead of snmplib/snmpIPBaseDomain.c
such that it is available in all Net-SNMP source files when compiling
Net-SNMP on Solaris ?
I've followed the approach that was already used in
lanas wrote:
The -e option in trapsess (snmpd.conf) sets the contextEngineID. This
is easily observable using tcpdumps. The -E option in trapsess also
sets the contextEngineID. This goes against the snmpcmd manual page
where -e sets the authoritative security engineID. There is no way to
Tommy Beadle wrote:
I would like to propose the attached patch to the Python bindings. This is a
patch to version 5.4.2.1. Currently, including null-bytes in octet strings
doesn't work because it's just treating the strings from Python as
null-terminated strings instead of getting their
Thomas Anders wrote:
Do we have a problem with the test scripts or with trap sending over a
streamed transport itself?
Looks like a problem with the test scripts only. Hopefully fixed in r18093 +
r18094.
+Thomas
Running
make test TESTOPTS=-P tcp
in trunk (and probably the branches, too) fails for some of the trap sending
tests:
27: testing snmpv1 traps are sent by snmpd API... FAIL
28: testing snmpv2c traps are sent by snmpd API... FAIL
29: testing snmpv1 traps are sent by snmpd without notification
The current changes in SVN trunk (5.6-to-be) seem to have introduced build
failures on quite a few systems:
- Solaris 2.6 + 7, HP-UX 10.20 + 11.00 with cc:
/my/net-snmp-5.6svn/snmplib/snmpUDPDomain.c, line 1115: error 1588:
INET_ADDRSTRLEN undefined.
- Solaris 8 + 9 + 10 with cc:
Ceylan, Hasan wrote:
exec Pages_free /bin/sh /etc/snmp/cat.sh Pages_free
exec Pages_free /bin/sh /etc/snmp/cat.sh Pages_active
exec Pages_free /bin/sh /etc/snmp/cat.sh Pages_inactive
exec Pages_free /bin/sh /etc/snmp/cat.sh Pages_wired_down
exec Pages_free /bin/sh /etc/snmp/cat.sh
r17933 introduced an additional #include net-snmp/library/types.h in
include/net-snmp/library/tools.h.
configure now throws the following #error for me:
- --- snip ---
In file included from /my/net-snmp-5.6svn/include/net-snmp/library/types.h:9,
from
bvass...@users.sourceforge.net wrote:
Revision: 17913
http://net-snmp.svn.sourceforge.net/net-snmp/?rev=17913view=rev
Author: bvassche
Date: 2009-12-17 13:05:53 + (Thu, 17 Dec 2009)
Log Message:
---
Converted starttime from a global variable into a static
Jordan Kagan wrote:
If I have questions on developing new capabilities and modifying
net-snmp source code, is this the correct address to send those
inquiries to?
Sure. Volunteers welcome.
Patches should be submitted to the patch tracker:
http://www.net-snmp.org/patches .
+Thomas
Max Romanov wrote:
It's me again. I've checked the trunk source, and found the issue fixed.
I'm apologize for false alarm and would like to suggest another small fix:
Again, please submit to http://www.net-snmp.org/patches so it won't get lost.
+Thomas
Sayan Chaliha wrote:
The _snmp_sess_asynch_send() function deallocates the PDU allocated by
the snmp_create_pdu call.
In the context of Windows, there is a major drawback to this:
1. snmp_create_pdu does not allocate memory for PDU enterprise, and
community; the user application
Sayan Chaliha wrote:
undefined reference to netsnmp_transport_open_client
This symbol is in libnetsnmpagent. Either you didn't properly link with it or
you linked with an older version of it (that may live somewhere else on your
system).
+Thomas
Jan Safranek wrote:
sure, here is second version of the patch, now with
NETSNMP_NO_LEGACY_DEFINITIONS
+1 (wasn't that easy? :-))
+Thomas
--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
Jan Safranek wrote:
I don't think that suggested NETSNMP_NO_LEGACY_DEFINITIONS is good idea -
the application would see the old identifiers by default. IMHO it is better
to advertise the new ones and show the old only if application (or
its packager) explicitly ask so.
At the very least,
Jan Safranek wrote:
While testing net-snmp-rc1 I have noticed C++ packages (like tog-pegasus)
do not like include/net-snmp/library/types.h.
The file contains closing brace for the C++ guard, but not the opening one,
so I'd suggest to add one in this patch.
+1
+Thomas
1 - 100 of 1040 matches
Mail list logo
