Hi, is somebody actually maintaining the SNMP module? I'm currently using PHP 4.2.3 and I have two major problems which make the SNMP functions unusable for me: - When using snmpwalkoid() the array keys are the OIDs. But if the OID is lying under certain subtrees (e.g. mib-2, experimental, private), all the OID bits leading to that path are dropped (i.e. the OID is shortened). Furthermore known OID numbers are replaced with their textual representation, which makes the array keys dependent on the locally existing MIBs. All this makes it impossible to reliably parse the returned OIDs! A PHP function which calls ds_toggle_boolean(DS_LIBRARY_ID, DS_LIB_PRINT_NUMERIC_OIDS); and ds_toggle_boolean(DS_LIBRARY_ID, DS_LIB_PRINT_FULL_OID); to turn these "features" off is required. And IMHO this should be the default behaviour. - Returned OID values are read via the sprint_value() SNMP API function. This preparses the values and results in strings like "Timeticks: ...", "Hex: ...". It is extremely hard (read: impossible) to reliably parse these values for further processing. This behaviour could also be turned off with ds_toggle_boolean() calls, but it's even better to do the parsing in a special function. This way, even the type of the value could be returned optionally.
Two other things in the related PHP source code might be regarded as... ermmm... questionable: - sprint_value() and sprint_objid() are used. It seems to me that these functions don't do any overflow checking? And they are not documented, thus probably obsolete (sprint_variable() is documented as obsolete in the mib_api man page). - Is it really wise to use ONE function for ALL SNMP operations (get, set, walk)? The code is hard to read and confusing. IMHO the function should at least be splitted up into a single-OID-function (set, get) and a multiple-OIDs-function (walk). Comments? Cheers, Jonny <[EMAIL PROTECTED]> -- PHP Development Mailing List <http://www.php.net/> To unsubscribe, visit: http://www.php.net/unsub.php
