On Tue, May 27, 2008 at 5:42 PM, Emmanuel Lecharny <elecharny@apache.org> wrote:
Stefan Seelmann wrote:
Hi dev,

We are using the shared-ldap schema parser now in the LDAP Browser
plugin of Studio.

Some LDAP servers however have a schema that isn't valid according to
RFC's (for example the Netscape successors are using "nsFooBar-oid" as
OID for some  attributeTypes and objectClasses). This causes problems
when loading the schema from such servers as the schema parser is quite
strict.

I would like to add a flag to the schema parser to make the parser
strict or more relaxed, depending on the usage. The default behaviour
should be unchanged of course, we need a strict parser in the server and
in the schema editor.

Yeah sounds good to me too. 
 


 
Hi Stefan,

this makes perfect sense to me. We also have to relax the parser in many other aspects :
- allowing tabs instead of spaces,
- allowing more than one space
- allowing missing spaces before or after '(' and ')'
- allowing unordered parameters.

Also some case invariance might be a good idea.  The parser seems to blow up when there's mixed case: i.e. attributetype passes but not attributeType.
 

This will render the grammar a little (lot?) more complex, but this is really important. Many users have problems with the syntax checker right now.

Oh yeah but those are things we have to perhaps allow users to toggle on and off.  Like being able to relax certain kinds of schema checks per syntax, per attributeType etc.  Some low hanging fruit for those interested in getting involved in ADS. 
 

So it's up to you. I don't even think we need a flag for that, or may be a 'strict' flag to enforce strict syntax checking, the relaxed one being the default, IMO.