perl-ldap-dev

Re: Problem found: Problems with schema changes

[Jim H. <ha...@us...>]

Graham, Clif,

I agree that the entry is extremely difficult to parse.  We just
upgraded to a new version of Novell NDS (8.5) a week before Clif's
patches were circulated.  We were able to grab the schema the day we
upgraded with the unpatched Net::LDAP:Schema, but the combination of
patches and NDS upgrade broke things.  We don't use this for anything
critical at the moment.  I will bring this issue to Novell's attention.
Thanks for yours.

  --Jim Harle

On Wed, 7 Feb 2001, Graham Barr wrote:

> OK, the problem is an entry in your schema that the parser does
> not like. It is
>
>
> Cannot parse [( stgroupsimin-oid NAME 'sTGROUPSIMIN' DESC 'Standard
> Attribute' SYNTAX 1.3.6.1.4.1.1466.115.121.1.40{64512} SINGLE-VALUE
> X-NDS_NAME 'ST GROUPS I'M IN' X-NDS_NOT_SCHED_SYNC_IMMEDIATE '1' )] '
> ) at lib/Net/LDAP/Schema.pm line 493, <STDIN> chunk 1.
>
> The value after X-NDS_NAME contains a '   is that legal ? If so how
> do you tell the difference between the delimiting quotes and those within
> a value ?
>
> Graham.
>

Re: Problem found: Problems with schema changes

[Clif H. <cl...@di...>]

Jim and Graham,

What I find strange is why the schema code could parse the record
before the patches and not afterwords.  I/we did not change anything
in the parse code.

I also think that the ' in I'M should be escaped with a
\ if it is going to be used this way.

Clif Harden


> 
> Graham, Clif,
> 
> I agree that the entry is extremely difficult to parse.  We just
> upgraded to a new version of Novell NDS (8.5) a week before Clif's
> patches were circulated.  We were able to grab the schema the day we
> upgraded with the unpatched Net::LDAP:Schema, but the combination of
> patches and NDS upgrade broke things.  We don't use this for anything
> critical at the moment.  I will bring this issue to Novell's attention.
> Thanks for yours.
> 
>   --Jim Harle
> 
> On Wed, 7 Feb 2001, Graham Barr wrote:
> 
> > OK, the problem is an entry in your schema that the parser does
> > not like. It is
> >
> >
> > Cannot parse [( stgroupsimin-oid NAME 'sTGROUPSIMIN' DESC 'Standard
> > Attribute' SYNTAX 1.3.6.1.4.1.1466.115.121.1.40{64512} SINGLE-VALUE
> > X-NDS_NAME 'ST GROUPS I'M IN' X-NDS_NOT_SCHED_SYNC_IMMEDIATE '1' )] '
> > ) at lib/Net/LDAP/Schema.pm line 493, <STDIN> chunk 1.
> >
> > The value after X-NDS_NAME contains a '   is that legal ? If so how
> > do you tell the difference between the delimiting quotes and those within
> > a value ?
> >
> > Graham.
> >
> 
> 
> 
> 


-- 

Regards,

Clif Harden           INTERNET:  c-h...@ti...
Texas Instruments                
Directory Services
6500 Chase Oaks Blvd, M/S 8412  
Plano, TX 75023        
                         Voice:  972-575-0855
                           FAX:  972-575-2418

Re: Problem found: Problems with schema changes

[John B. <joh...@ne...>]

Hi folks.

Sorry for trouble with the clearly-not-quite-production schema code.

On Thu, 8 Feb 2001, Clif Harden wrote:

> What I find strange is why the schema code could parse the record
> before the patches and not afterwords.  I/we did not change anything
> in the parse code.
>
> I also think that the ' in I'M should be escaped with a
> \ if it is going to be used this way.

Well, you (and I :-) would like to think so. I may be way off, but I don't
see the standard as being clear.

from RFC2252:

"Terms which begin with the characters "X-" are reserved for private
experiments, and MUST be followed by a <qdstrings>."

and:

"utf8 = <any sequence of octets formed from the UTF-8 [9]
                       transformation of a character from ISO10646 [10]>

   dstring = 1*utf8

    qdstring = whsp "'" dstring "'" whsp

    qdstringlist = [ qdstring *( qdstring ) ]

    qdstrings = qdstring / ( whsp "(" qdstringlist ")" whsp )"


So...why can't qdstrings contain unquoted ' and ) chars?

What have I missed?


jb

Re: Problem found: Problems with schema changes

[Graham B. <gb...@po...>]

I went a re-read the RFC too, and scratched my head. Personally I think
it must be an oversight in the RFC.

Graham.

On Thu, Feb 08, 2001 at 03:49:20PM +0000, John Berthels wrote:
> 
> Hi folks.
> 
> Sorry for trouble with the clearly-not-quite-production schema code.
> 
> On Thu, 8 Feb 2001, Clif Harden wrote:
> 
> > What I find strange is why the schema code could parse the record
> > before the patches and not afterwords.  I/we did not change anything
> > in the parse code.
> >
> > I also think that the ' in I'M should be escaped with a
> > \ if it is going to be used this way.
> 
> Well, you (and I :-) would like to think so. I may be way off, but I don't
> see the standard as being clear.
> 
> from RFC2252:
> 
> "Terms which begin with the characters "X-" are reserved for private
> experiments, and MUST be followed by a <qdstrings>."
> 
> and:
> 
> "utf8 = <any sequence of octets formed from the UTF-8 [9]
>                        transformation of a character from ISO10646 [10]>
> 
>    dstring = 1*utf8
> 
>     qdstring = whsp "'" dstring "'" whsp
> 
>     qdstringlist = [ qdstring *( qdstring ) ]
> 
>     qdstrings = qdstring / ( whsp "(" qdstringlist ")" whsp )"
> 
> 
> So...why can't qdstrings contain unquoted ' and ) chars?
> 
> What have I missed?
> 
> 
> jb
> 
> 
> 
> 
> 

Re: Problem found: Problems with schema changes

[Kurt D. Z. <Ku...@Op...>]

At 05:12 PM 2/8/01 +0000, Graham Barr wrote:
>I went a re-read the RFC too, and scratched my head. Personally I think
>it must be an oversight in the RFC.

I concur.  I also not that whsp is defined as ZERO or more
spaces.... fun, eh?

Kurt

Re: Problem found: Problems with schema changes

[Graham B. <gb...@po...>]

On Thu, Feb 08, 2001 at 09:42:14AM -0800, Kurt D. Zeilenga wrote:
> At 05:12 PM 2/8/01 +0000, Graham Barr wrote:
> >I went a re-read the RFC too, and scratched my head. Personally I think
> >it must be an oversight in the RFC.
> 
> I concur.  I also not that whsp is defined as ZERO or more
> spaces.... fun, eh?

Yes, if it wa one or more the we could at least determine this
particular case as it had an alpha character on either side.

Graham.

Re: Net::LDAP::bind problems under Apache

[Gary Ashton-J. <gas...@f2...>]

I haven't fixed the problem but I have identified the circumstances which
cause Net::LDAP::bind to always fail under Apache.

I accidentally left a PerlAuthenHandler Apache::AuthNetLDAP in one
<Location> while testing my variation of AuthTicket which uses NetLDAP in
another <Location>. Calls to bind were always failing. Eventually I realised
the clash, commented out the AuthNetLDAP call and it works beautifully.

Thanks Graham Barr for the patch to ASN1.pm which stops the method
redefinition warnings but I think now the patch might be better left out.
They were obviously correct warnings because Net::LDAP was being called
twice and I should have been willing to look further.

OTOH maybe Net::LDAP needs to handle(!) being called by two different
handlers...

Rgds

Gary

Re: Problem found: Problems with schema changes

[Chris R. <chr...@me...>]

Clif Harden <cl...@di...> wrote:
> 
> Jim and Graham,
> 
> What I find strange is why the schema code could parse the record
> before the patches and not afterwords.  I/we did not change anything
> in the parse code.
> 
> I also think that the ' in I'M should be escaped with a
> \ if it is going to be used this way.
> 
> Clif Harden

RFC 2252 doesn't say how to deal with this. Incidentally this issue has
just come up on the ldapbis mailing list, and there are a couple of
suggestions on how to fix it - \hex, \char, or doubling-up-the-quote eg ''.
(I think \hex is more like the quoting used in other bits of LDAPv3 and
therefore more sensible to use. Other peoples' mileage varies. Handling
doubling-up in a regex is a bit of a pain.)

I'd say it is a pretty dubious thing to have in a schema token, and the
vendor should certainly be notified (which Jim's done.)

Cheers,

Chris