"This has been done a couple of times." Is someone using iBatis in conjunction with LDAP? I am aware of our previous discussions. But, where is the implementation?
Brandon On 2/12/06, Clinton Begin <[EMAIL PROTECTED]> wrote: > > This has been done a couple of times. It was discussed on this list a > while back. > > Cheers, > Clinton > > Cimballi wrote: > > >Great ! It's more motivating to work when you know other people may > >enjoy it ! ;-) > > > >I hope language will not block me, my english is very poor ! Well, at > >least I understand "computer" english, and I've got a big dictionnary > >! > > > >So Larry, like you said in LDAP each attribute can have multiple > >values. This is one of the differences between LDAP and SQL. If I > >understand you well, you also propose a kind of automatic mapping > >between the object (JAVA) and the entry (LDAP) based on the > >type-objectclass of that object-entry. I don't understand well your > >reference to SQL injection, but maybe this is because I don't still > >have a look inside the sqlmap architecture. > > > >Now what I thought to do. The simplest way to start should to copy the > >way sqlmap works and adapt it to LDAP. Of course there is some > >difference but we will discuss them later, I don't think they should > >need a new architecture. > >About the about to use, I've mostly work with the ldapsdk from > >netscape-sun-mozilla, but maybe it's should be better to use JNDI > >here. About JNDI, do you think it should be interesting to have a > >generic JNDI dao engine, or maybe we can start with JNDI-LDAP dao > >engine, and look for more genericity later ? > >About mapping between JAVA and LDAP : in LDAP most of the attributes > >can be multi valued, but in real life, company's schemas usually > >settled if an attribute will be mono or multi valued based on their > >model, so maybe we can provide a parameter in the mapping which will > >say if the engine should consider the attribute as mono or multi > >valued. > >About transactions, although there is no transaction in LDAP for the > >moment, we should provide the same functionnely like in sqlmap so it > >could be easily implemented later. > >This is what I've got in mind for now. > > > >I'd like to know about exchange of docs and code inside iBatis > >community. If I'd like to submit you docs or code, what's the best way > >? Should I post them on this list ? > > > >If I use some words that don't match the ones you are used to use, > >just tell me ! > > > >First step is made, hope it was the most difficult one ;-) > > > >Cimballi > > > > > >On 2/12/06, Larry Meadors <[EMAIL PROTECTED]> wrote: > > > > > >>I think that would be very interesting. :-) > >> > >>I did some work with LDAP recently, and think that having something > >>like SQL Maps for LDAP queries would be really useful. One thing I had > >>problems with was that the attributes are all name-value pairs - like > >>doing everything with Maps. ;-) > >> > >>Semantically, I think it would fit - there is no concept of a mapped > >>statement, but SQL injection is not really an issue, so escaping may > >>be adequate. Query results come back in a similar way (name-value > >>pairs), so that fits. It may actually be simpler because the names of > >>the attributes are defined by the classes of the returned objects. > >> > >>Sounds cool! > >> > >>Larry > >> > >> >