Having some issues moving subagent from running with 5.8 to 5.9

--===============8971373869738358770==
Content-Type: multipart/alternative; boundary="00000000000018bd78061f808d64"

--00000000000018bd78061f808d64
Content-Type: text/plain; charset="UTF-8"

I am trying to get a subagent to work after moving from net-snmp 5.8 to
5.9, but it keeps segfaulting for me.  I have tracked down that it is
failing when trying to sort the table due to the structure not being what
it should be.  It is using table_container to store the tables that are
causing issues.  If I switch them to linked_list, they work fine, and no
more bad pointer accesses.  The bad pointer accesses are showing as trying
to compare oids where the length is in fact one of the pointers rather than
the actual length of the oid, so it makes sense to be crashing when that
happens.  The code has not really changed in years and has been working
with many versions of net-snmp so far, but 5.9 for reasons I have failed to
figure out is different in some way.  Nothing in the changelogs or diff'ing
the header files have pointed me in the right direction.  I don't think
switching to linked_list is the right thing to do since that might make it
quite a bit slower.

Does anyone have any idea where I should have a look to try and narrow down
where the data structure is ending up wrong when using table_container (aka
binary_array)?

Thanks.

-- 
Len

--00000000000018bd78061f808d64
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div>I am trying to get a subagent to work after moving fr=
om net-snmp 5.8 to 5.9, but it keeps segfaulting for me.=C2=A0 I have track=
ed down that it is failing when trying to sort the table due to the structu=
re not being what it should be.=C2=A0 It is using table_container to store =
the tables that are causing issues.=C2=A0 If I switch them to linked_list, =
they work fine, and no more bad pointer accesses.=C2=A0 The bad pointer acc=
esses are showing as trying to compare oids where the length is in fact one=
 of the pointers rather than the actual length of the oid, so it makes sens=
e to be crashing when that happens.=C2=A0 The code has not really changed i=
n years and has been working with many versions of net-snmp so far, but 5.9=
 for reasons I have failed to figure out is different in some way.=C2=A0 No=
thing in the changelogs or diff&#39;ing the header files have pointed me in=
 the right direction.=C2=A0 I don&#39;t think switching to linked_list is t=
he right thing to do since that might make it quite a bit slower.</div><div=
><br></div><div>Does anyone have any idea where I should have a look to try=
 and narrow down where the data structure is ending up wrong when using tab=
le_container (aka binary_array)?</div><div><br></div><div>Thanks.<br></div>=
<div><div><br><span class=3D"gmail_signature_prefix">-- </span><br><div dir=
=3D"ltr" class=3D"gmail_signature" data-smartmail=3D"gmail_signature">Len</=
div></div></div></div>

--00000000000018bd78061f808d64--


--===============8971373869738358770==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline


--===============8971373869738358770==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
Net-snmp-coders mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/net-snmp-coders

--===============8971373869738358770==--