Re: inetmib1(3/31): Add tests for SnmpExtensionInit

[Paul Vriens]

Juan Lang wrote:
> --Juan

These tests import snmpapi, which isn't available on win95. Do we need to cater 
for that and do LoadLibrary/GetProcAddress ?

inetmib1(3/31): Add tests for SnmpExtensionInit

[Juan Lang]

> These tests import snmpapi, which isn't available on win95. Do we need to
> cater for that and do LoadLibrary/GetProcAddress ?

Hm.  Yeah, I suppose so.  That or duplicate the SnmpUtil functions the
tests use, which shouldn't be that hard either.
--Juan

inetmib1(3/31): Add tests for SnmpExtensionInit

[Alexandre Julliard]

"Juan Lang"  writes:
> > These tests import snmpapi, which isn't available on win95. Do we need to
> > cater for that and do LoadLibrary/GetProcAddress ?
>
> Hm.  Yeah, I suppose so.  That or duplicate the SnmpUtil functions the
> tests use, which shouldn't be that hard either.

Is there really a Win95 setup that has inetmib1 but not snmpapi?

inetmib1(3/31): Add tests for SnmpExtensionInit

[Paul Vriens]

Alexandre Julliard wrote:
> "Juan Lang"  writes:
> > > These tests import snmpapi, which isn't available on win95. Do we need to
> > > cater for that and do LoadLibrary/GetProcAddress ?
> >
> > Hm.  Yeah, I suppose so.  That or duplicate the SnmpUtil functions the
> > tests use, which shouldn't be that hard either.
>
> Is there really a Win95 setup that has inetmib1 but not snmpapi?

Yes, mine :-) and from the looks at 
http://test.winehq.org/data/92c8cac214715633459227b2d47df61a18dc8d82/ so does 
Francois' box.

inetmib1(3/31): Add tests for SnmpExtensionInit

[Juan Lang]

> > These tests import snmpapi, which isn't available on win95. Do we need to
> > cater for that and do LoadLibrary/GetProcAddress ?
>
> Hm.  Yeah, I suppose so.  That or duplicate the SnmpUtil functions the
> tests use, which shouldn't be that hard either.

On second thought, I don't see how the tests can succeed without
snmpapi.  In particular, SnmpExtensionQuery sure seems to call
SnmpUtilOidAppend with called with SNMP_PDU_GETNEXT, in which case the
caller should call SnmpUtilOidFree.

We could either leak memory in order to avoid depending on snmpapi,
and duplicate SnmpUtilOidNCmp and SnmpUtilOidToA in the tests, or skip
the inetmib1 tests when snmpapi isn't available.
--Juan