Rich Adamson
2006-Apr-01 16:17 UTC
[Asterisk-Users] Install problem with res_snmp.so from current trunk (bug?)
Just updated two fc3 systems running svn trunk. One updated, installed properly, and is working fine. The second box failed during the 'make install' process with: /usr/lib/libnetsnmp.a(parse.o)(.text+0x275a): In function `unload_module': : multiple definition of `unload_module' res_snmp.o(.text+0x310):/usr/src/asterisk/res/res_snmp.c:102: first defined here /usr/bin/ld: Warning: size of symbol `unload_module' changed from 66 in res_snmp.o to 284 in /usr/lib/libnetsnmp.a(parse.o) collect2: ld returned 1 exit status make[1]: *** [res_snmp.so] Error 1 make[1]: Leaving directory `/usr/src/asterisk/res' make: *** [subdirs] Error 1 I noticed in the res/Makefile includes the following: # NETsnmp has some difficulties on some platforms (conflict with unload_module) # Until we figure out if the collission is version-specific or what to do # we have disabled res_snmp on OS/X and *BSD The two fc3 boxes have different versions of the /usr/lib/libnetsnmp modules (since they are different sizes on the two boxes). My work around was to simply comment out the res/Makefile steps to compile the res_snmp module, and to use 'noload res_snmp.so'. Is this worthy of opening a bug assuming the above comment is still valid? Would the individual(s) maintaining res_snmp want to log into either of these internet accessible boxes to identify the root cause? Rich
Kevin P. Fleming
2006-Apr-01 16:24 UTC
[Asterisk-Users] Install problem with res_snmp.so from current trunk (bug?)
Rich Adamson wrote:> Is this worthy of opening a bug assuming the above comment is still > valid? Would the individual(s) maintaining res_snmp want to log into > either of these internet accessible boxes to identify the root cause?The module loader in trunk is undergoing changes that will eliminate this problem very soon.
Rich Adamson
2006-Apr-01 20:27 UTC
[Asterisk-Users] Install problem with res_snmp.so from current trunk (bug?)
Kevin P. Fleming wrote:> Rich Adamson wrote: > >> Is this worthy of opening a bug assuming the above comment is still >> valid? Would the individual(s) maintaining res_snmp want to log into >> either of these internet accessible boxes to identify the root cause? > > The module loader in trunk is undergoing changes that will eliminate > this problem very soon.No problem. Thanks.