- Libsmi - lists.ibr.cs.tu-bs.de

SMIv2 pretty print
by Andy Bierman 30 Apr '03

30 Apr '03

Hi, I haven't looked at the code yet, but... I wanted to check with the smidump authors about the feasibility of modifying the smidump source code to act as an SMIv2 pretty print tool. I realize the complexity of such a tool depends on how configurable it is. I don't need any user configuration. I noticed that smidump -f smiv2 will do some pretty printing, such as changing {foo 2} to { foo 2 }. I need this tool to format MIBs as they usually appear in RFCs: - preserve ASN.1 comments - one enum per line in INTEGER objects or TCs - one descriptor and type per line in SEQUENCEs - one descriptor per line in OBJECTS and OBJECT-GROUP clauses - uniform indentation and whitespace - smart reformat to deal with lines > 72 chars - many more details such as these? I would be willing to work on the code, but I wanted to check first with this group if this is possible or if anyone else has thought about such a tool. thanks, Andy

2 1

[Fwd: [libsmi] problem restricting TimeTicks - bad error message???]
by Pete Flugstad 30 Apr '03

30 Apr '03

I didn't see any response from this? In summary: A) is restricting TimeTicks like the below, legal SMIv2? testInteger OBJECT-TYPE SYNTAX TimeTicks (500..12000) B) even if it's not, the error message libsmi spits out when it sees this is totally wrong and misleading: the SPPI construct/keyword `MAX-ACCESS' may not be used in a MIB Pete Flugstad Icon Labs

3 3

smilint bug for AGENT-CAPABILITIES
by Andy Bierman 29 Apr '03

29 Apr '03

Hi, I am getting the following error message for a valid AGENT-CAPABILITIES file, related to a DEFVAL clause: ./CISCO-CAS-IF-CAPABILITY.my.1.0:74: [2] default value syntax does not match object syntax snippet from source file: VARIATION ccasVoiceCfgNonLinearProcEnable DEFVAL { true } DESCRIPTION "The initial value is 'true'." The object is defined in the CISCO-CAS-IF-MIB as SYNTAX TruthValue. program version: smilint 0.4.1 Andy

2 2

Re: smidump -f smiv1 output
by Kristine Adamson 02 Apr '03

02 Apr '03

Frank/Juergen, Thanks for the explanation. We had been using a different program to convert our SMIv2 MIB module to SMIv1 prior to using smidump. This other program used to put the SMIv1 definition of the textual conventions inline in the SMIv1 output, so we expected libsmi to do the same. Thanks for clearing up our confusion, Kristine Adamson IBM Communications Server for MVS: TCP/IP Development Internet e-mail:adamson@us.ibm.com

1 0

Re: [libsmi] Re: smidump -f smiv1 output
by Kristine Adamson 02 Apr '03

02 Apr '03

Hi! Frank Strauss writes: Frank>Such a change is not necessary. The SMIv1 output as it is produced now Frank>is valid and reflects more closely what the original SMIv2 defines. What was produced as output was the same definition as existed in the SMIv2 MIB module. Will a MIB browser that only understands SMIv1 know what an InetAddressType SYNTAX is? Kristine Adamson IBM Communications Server for MVS: TCP/IP Development Internet e-mail:adamson@us.ibm.com

3 2

Re: smidump -f smiv1 output
by Kristine Adamson 02 Apr '03

02 Apr '03

Hi and thanks for the feedback. For example, we have the following SMIv2 MIB object: ibmMvsTcpListenerLocalAddrType OBJECT-TYPE SYNTAX InetAddressType MAX-ACCESS read-only STATUS current DESCRIPTION "The IP address type of the ibmMvsTcpListenerLocalAddr object." ::= { ibmTcpipMvsTcpListenerEntry 2 } In the SMIv1 output from smidump, we thought this object would have been defined as: ibmMvsTcpListenerLocalAddrType OBJECT-TYPE SYNTAX INTEGER { unknown(0), ipv4(1), ipv6(2), ipv4z(3), ipv6z(4), dns(16) } ACCESS read-only STATUS mandatory DESCRIPTION "The IP address type of the ibmMvsTcpListenerLocalAddr object." ::= { ibmTcpipMvsTcpListenerEntry 2 } Thanks, Kristine Adamson IBM Communications Server for MVS: TCP/IP Development Internet e-mail:adamson@us.ibm.com

2 1

smidump -f smiv1 output
by Kristine Adamson 02 Apr '03

02 Apr '03

Hello, We are trying to use the -f smiv1 option of smidump to create an SMIv1 version of an SMIv2 MIB module. In the SMIv1 output version, we expected that the SYNTAX of those MIB objects defined with textual conventions, would have been changed from the textual convention to the corresponding base SMIv1 syntax type. But none of the textual conventions were updated in the SMIv1 output version. Is our understanding correct of the -f smiv1 option on smidump? Thanks, Kristine Adamson IBM Communications Server for MVS: TCP/IP Development Internet e-mail:adamson@us.ibm.com

2 1

(no subject)
by Steve Bucknam 01 Apr '03

01 Apr '03

Is there a reason why the folloing line is commented out for the indexes in the dump-xsd.c file? Now most infomation is missing, such as oid. fprintAnnotationElem( f, smiNode ); Steve

2 1

problem restricting TimeTicks - bad error message???
by Pete Flugstad 31 Mar '03

31 Mar '03

I have encountered a mib (full mib attached) with the following entry: testInteger OBJECT-TYPE SYNTAX TimeTicks (500..12000) MAX-ACCESS read-write STATUS current DESCRIPTION "Test Integer" ::= { testObjects 1 } When I run this through libsmi (0.4.1), smilint, I get: test.mib:22: the SPPI construct/keyword `MAX-ACCESS' may not be used in a MIB and it bails. It's pointing at the SYNTAX line. Now, I don't claim to be a SMIv2 expert, but I believe the syntax is OK, isn't it? Maybe not? Even if it's not, the error message is completely wrong and misleading, as the problem has nothing at all to do with the MAX-ACCESS line. Pete Flugstad Icon Labs TEST-MIB DEFINITIONS ::= BEGIN IMPORTS MODULE-IDENTITY, OBJECT-IDENTITY, NOTIFICATION-TYPE, TimeTicks, OBJECT-TYPE, enterprises FROM SNMPv2-SMI; testMibModule MODULE-IDENTITY LAST-UPDATED "200301230000Z" ORGANIZATION "World" CONTACT-INFO "Someplace" DESCRIPTION "Test module" ::= { enterprises 5687 } --- comment test testObjects OBJECT IDENTIFIER ::= { testMibModule 1 } testInteger OBJECT-TYPE SYNTAX TimeTicks (500..12000) MAX-ACCESS read-write STATUS current DESCRIPTION "Test Integer" ::= { testObjects 1 } testObjectIdentity2 OBJECT-IDENTITY STATUS current DESCRIPTION "Test Object Identity 2" ::= { testObjects 2 } testObjectIdentity3 OBJECT-IDENTITY STATUS current DESCRIPTION "Test Object Identity 3" ::= { testObjects 3 } testAlarms OBJECT IDENTIFIER ::= { testMibModule 2 } testAlarmIdentity1 OBJECT-IDENTITY STATUS current DESCRIPTION "Test Alarm Identity 1" ::= { testAlarms 1 } testAlarmIdentity2 OBJECT-IDENTITY STATUS current DESCRIPTION "Test Alarm Identity 2" ::= { testAlarms 2 } testAlarmIdentity3 OBJECT-IDENTITY STATUS current DESCRIPTION "Test Alarm Identity 3" ::= { testAlarms 3 } testTraps OBJECT IDENTIFIER ::= { testMibModule 3 } testTrapPrefix OBJECT IDENTIFIER ::= { testTraps 0 } testTrap NOTIFICATION-TYPE OBJECTS { testInteger } STATUS current DESCRIPTION "test trap" ::= { testTrapPrefix 1 } END

1 0

Traversing nodes
by Garth R Conrad 19 Mar '03

19 Mar '03

Hi, I'm new to libsmi and have a question that is probably easy but is baffling me. I've written some code that deals with a single mib module ("A"). I want to traverse all the nodes in depth-first order (or "tree pre-order", as the documentation says). I use smiLoadModule(), followed by smiGetFirstNode() and smiGetNextNode() to loop through the nodes. This works as expected - I see the nodes defined in A in depth-first order. Now I'm trying to generalize my code to load N modules at program startup. After loading all the modules, is there a way to loop through all the nodes defined in _any_ of those modules so that the nodes are returned in depth-first order? If so, how? Any feedback would be much appreciated. Thanks, Garth Conrad IBM Communications Server for MVS

2 1