Modules this distribution does not carry¶
Two different things, on one page because a contributor offering either gets the same answer.
Removed: 31 modules index-frozen.csv names and this distribution no
longer carries, with the reason each was dropped. That is the bulk of the page
and everything below down to the count.
Declined: 13 modules another collection ships that this distribution was offered and turned down. They were never here, so the snapshot never named them and no index row points at them; what they have in common with the removed ones is only that the answer to “should we carry this?” is already written down.
index-frozen.csv is a snapshot of what index.csv answered before this
repository began correcting it, and it is never rewritten: an OID that had an
answer keeps that answer. That guarantee covers the rows, not the modules they
name. 31 of the modules named there are carried neither in src/ nor in
pysmi’s bundle, so asn1/NAME and json/NAME.json return 404 for them, and
223 index rows name a module this site does not serve.
Nothing fails to compile because of this. No module still in the corpus imports any of the 31. What is lost is the site answering for their OIDs, and most of these are leaf modules, which are polled directly and imported by nothing. That nothing imports a module is not evidence that nothing used it.
This page is the record of which modules those are and why each was dropped.
modules named by |
5,347 |
carried nowhere |
31 |
index rows naming them |
223 |
surviving modules with a dangling import |
0 |
What used to be here¶
Two groups have gone, and they were the bulk of the page: 65 consortium
modules and 98 RFC modules that pysmi’s manifest named but its wheel did not
ship, held in pysmi/mibs/future/. This repository had deleted its own copies
in #351, #375 and #384 on the grounds that the bundle supplied them, and then
the bundle stopped.
pysnmp/pysmi#323 promoted the whole held tier – all 274 modules, 163 of them named here – so the wheel now carries every module its manifest names. The hold had rested on a survey showing nothing in this corpus imported one of them, which was true and was the wrong question to ask of a general-purpose compiler. What settled it was that pgmillon/observium, an independent monitoring distribution, ships 200 of the 274.
Every index row that named one of them resolves again. What is left below is the group that was never about packaging.
Deleted here deliberately: 31 modules¶
These are not an accident of packaging. Each was removed by a decision recorded in the PR named beside it, and none of them should come back.
why |
modules |
|---|---|
Obsolete. Publisher marked every object |
AC-PM-ATM-MIB, AcAtm, ARUBAWIRED-MCLAG-MIB, CISCO-GPRS-GTP-MIB, CTRON-SSR-L2-MIB, CTRON-SSR-L3-MIB, HP-ENTITY-MIB, Juniper-RADIUS-Disconnect-MIB, VMWARE-VCOPS-EVENT-MIB |
Superseded. Pre-publication draft whose published successor is bundled, often under a different name (#358) |
ADSL-DMT-LINE-MIB, DOCS-BPI2-MIB, DOCS-CABLE-DEVICE-TRAP-MIB, DOCS-IF-EXT-MIB, DOCS-SUBMGT-MIB, INT-SERV-GUARANTEED-MIB, ISIS-D13-MIB, MPLS-TC-MIB, MSTP-MIB, RTCPXR-MIB, XGCP-MIB |
Superseded. Multicast drafts, importers repointed to the RFC (#367) |
DVMRP-MIB, IGMP-MIB, IPMROUTE-MIB |
Never published. Draft that never became a standard, squatting an assigned |
INT-SERV-MIB, SYSLOG-MIB, T11-FC-SP-CERTS-MIB, T11-ZONE-SERVER-MIB |
Withdrawn. RFC 8096 withdrew the module (pysmi#176) |
IPV6-ICMP-MIB, IPV6-TCP-MIB, IPV6-UDP-MIB |
Not bundled, no importers (#375) |
COFFEE-POT-MIB |
Declined: 13 modules¶
Modules pgmillon/observium ships that this distribution does not, where the reason is a decision rather than an oversight. Auditing observium’s whole tree against ours left fourteen standard modules carried nowhere; one of them went into pysmi’s bundle in pysnmp/pysmi#325 and these thirteen did not.
why |
modules |
|---|---|
Superseded. SMIv1 predecessor whose successor pysmi bundles, on the same arcs: |
RFC1229-MIB, RFC1231-MIB, RFC1243-MIB, RFC1354-MIB |
Superseded framework. The SNMPv2 party administration SNMPv2c and v3 abandoned, written against the 1993 draft SMI: SNMPv2-PARTY-MIB imports |
SNMPv2-PARTY-MIB, SNMPv2-M2M-MIB |
Experimental, unused. pysmi bundles an Experimental RFC module only while something imports symbols it alone defines; nothing here imports any of these. Its exclusion list names all five (pysmi#325) |
AGGREGATE-MIB, TIME-AGGREGATE-MIB, RSERPOOL-MIB, SMF-MIB, IANA-SMF-MIB |
Never published. No RFC publishes it, and it names no |
SAVI-MIB |
Not attributed. IEEE 802.1, which wants an |
LLDP-EXT-DOT1-PE-MIB |
The count reads declarations, not filenames¶
tests/index-contract.sh counts these by comparing the names in the frozen
index against what src/ and the bundle carry. Until this was written it
compared against file basenames, which is not the same question: a file may
declare more than one module. src/vendor/extreme/EXTREME-BASE-MIB declared 34,
33 of them with no file of their own, and all 34 compiled and served, so
asn1/EXTREME-VLAN-MIB returned 200 while the check counted those 33 as carried
nowhere.
The check reads the NAME DEFINITIONS ::= BEGIN line out of each file, so the
number it pins is the number above. src/ no longer holds a file like that one:
that MIB is stored as 34 files, and tests/source-layout-contract.py fails the
build if any file under src/ declares more or fewer than one module.