[Date Prev][Date Next]
Re: "Active" backends
Look in the archives and ITS for back-sock, it's probably the approach
that makes sense going forward. I don't know of anyone on the Project
spending any time with that code yet though.
I have lately been trying to implement an OpenLDAP server that generates
some of the information on the fly and have therefore been pestering
different knowledgeble persons connected to OpenLDAP as well as this
Whereas I can't claim to have come up with a great solution I think I
might at least be able to ask the right questions, so here goes:
I have read that both back-tcl (which I wasn't too attracted to anyway)
and back-shell are deprecated and people have been telling me that
back-perl is buggy. It has even been suggested on the devel list that this
perhaps too should be deprecated. In light of this, is there any
recommended way of generating fairly arbitrary data? In the documentation
it says that back-shell is mainly meant for prototyping, prototyping for
what (considering that writing a complete new backend is prohibitively
difficult and time consuming for 99.9% of the users)?
Further I wonder what the exact problem is with back-shell. This I got to
work, but if there are subtle problems I guess I might find out when I
least want to.
Has a "back-fork" been considered? Just a minimal backend that starts
(fork - exec) a new process without even invoking a shell? Perhaps this is
a silly idea.
Lastly I would like to emphasize that even though this mail might be
construed as a bit "grumpy" I appreciate that a the developers work hard
on this project and I am glad to have access to it.
-- Howard Chu
Chief Architect, Symas Corp. http://www.symas.com
Director, Highland Sun http://highlandsun.com/hyc
OpenLDAP Core Team http://www.openldap.org/project/