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

RE: Race condition in ldif2ldbm (ITS#140)

You're right, but moving the maxkids=cmdkids after the fork of
ldif2id2children fixes it.  Thank you for pointing that out.

-----Original Message-----
From: Dave Boyce [mailto:dave@chello.com]
Sent: Friday, May 07, 1999 6:06 AM
To: sasgwi@wnt.sas.com
Cc: openldap-its@OpenLDAP.org
Subject: Re: Race condition in ldif2ldbm (ITS#140)

sasgwi@wnt.sas.com wrote:
> I think this would do the same job:
> diff -r1.25 ldif2ldbm.c
> 60a61
> >     int     cmdkids = 1;
> 90c91
> <                       maxkids = atoi( optarg );
> ---
> >                       cmdkids = atoi( optarg );
> 171a173,174
> >
> >     maxkids = cmdkids;

(Sorry for taking so long to reply to this: I've been looking at other

The above doesn't work. In fact, it makes absolutely no difference to
the operation of ldif2ldbm. Think about what happens: id2entry gets
then cmdkids gets written with the same value it would have on startup,
so ldif2ldbm carries on exactly as before.

Please back out the above 'fix' from the development branch, because it
wasn't tested, and doesn't work.

	- Dave