Luis EG Ontanon wrote:
> libsmi (that is going to be the one) has a rich collection of MIBs.
> We could ask them whether or not we can redistribute them.
libsmi's MIB collection have been editted to correct syntactic errors.
libsmi's parser is MUCH more picky than net-snmp's parser. I ran the
complete Cisco MIB collection through smilint and got about 1800 lines
of errors. And that was at level 4 which only reports the more important
errors. Compared with many other vendors, CIsco's MIBs are reasonably clean.
It would make more sense to me to use something equivalent to:
smidump -f identifiers -o <MIB-NAME>.oid <MIB-NAME>
than have one file. It means the user only has to generate the oid
mapping for new MIBs, rather than rebuilding an entire database, at the
cost of having to read multiple files.
--
There's no point in being grown up if you can't be childish sometimes.
-- Dr. Who
