[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: ldapmodify "daemon"



Sounds a bit like slurpd(8)... :-)

This goes a bit beyond the intent of ldapmodify(1)...
I think I rather keep the tools "simple" wrappers
for the API routines (which are simple wrappers around
protocol operations).

Kurt


At 01:20 AM 12/9/2002, Hallvard B Furuseth wrote:
>I'm patching ldapmodify to have an option to run as a kind of daemon:
>For each LDIF record you write to its stdin, it responds immediately
>after the modify request completed with the operation's result code, its
>name, and the error string on one line to stdout.  Thus, a program can
>invoke ldapmodify, feed it LDIF records and react to the results, or a
>wrapper program can authenticate and encrypt the connection and run
>ldapmodify as a true "LDIF daemon".
>
>Is there general interest in such an option, or should I just make it a
>private hack?
>
>Can someone come up with a brief help text for usage() and suggest a
>manpage description?  It's wrong to call it a "daemon" since it is not
>listening to a port or anything, but I don't know what else to call it.
>
>An additional option: I imagine it could be useful to have the reject
>file be stdout, if ldapmodify is run as a daemon.  So I'll let '-S -'
>mean that failed LDIF records are appended to the initial status/error
>message.  With this option, each response will end with a blank line.
>
>-- 
>Hallvard