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

RE: No object found.....

Earlier today, Gremmen, Jeroen wrote:

> If I can load data in ldap but cannot search I must be doing something
> wrong because it's a stable ldap release and stable releases are known for
> working perfectly!

Somehow, I suspect you're just being sarcastic;  if, however, you're serious,
then you're somewhat mistaken as to the meaning of the "stable" label.

> Here I end my bogus analysis. The test suite can't check for everything,
> I'm the last one to argue that. But the test suite claims to do search
> operations (one of the most basic operations you can expect for ldap) and
> also claims it gets the correct results back.... WELL, IT DOESN'T!!!

Following on from your original message:

  "Wen you do a search on the ./tests/test-db/master.log file you will see
  '**** got 0 of 0' on every search!"

Having just built and tested OpenLDAP 1.2.7 in an attempt to recreate this
problem, I find that "./tests/test-db/master.log" doesn't contain any instance
of the phrase "got n of n" at all.  Could you perhaps provide some context for
this line?  My tests all pass such that upon completion, this phrase does not
appear in master.log

A search of the OpenLDAP source (for RE /got.*of/) shows the string:

  *** got %ld of %lu so far

in the file "./servers/slapd/connection.c";  according to the source, there
should be a line just before this one, along the lines of:

  ber_get_next on fd %d failed errno %d (%s)

At a guess, if this line appears before the "got n of n" in your master.log,
then the errno should give you some clue as to why the searches are failing.

> Kurt wrote:
> > If you do have further problems, be sure to post enough information
> > such that those you ask to help you will be able to do so.
> I don't have to because you have all the files yourself.... the entire
> test suite.

As above - clearly I'm not suffering the same fate as you with the test suite,
so you'd have to provide additional information (such as a few more lines from
master.log, without quoting the whole file).

> Two more remarks:
> - The test suite occasionally hangs where the 2nd thread is defunct.

This definitely doesn't happen for me;  I suspect it's due to the less than
stellar threads implementation in Linux (I'm using Solaris).  There has been
some talk about pre-emptive threads being a problem for OpenLDAP 1.x on
OpenLDAP mailing lists.

> > - I would like you to read the 1st reply I got to my posting:
> From: Andrew Davison [mailto:andrew@infradig.com] 
> Found the problem. Unfortunately i'll have to do a release for it. Soon
> as I tidy up some loose ends I will do so, later today hopefully.

With all due respect to Andrew Davison - I don't see his name as a core
OpenLDAP developer, and his presence seems pretty small in a search of all the
OpenLDAP mailing list archives;  just what is it that he plans to "release"?
Is there some other piece of software in which Andrew may have found a bug?

> Somebody seems to have found the problem.... Without blaming me for
> misusing the software!

Kurt didn't "blame" you for "misusing the software" - he merely stated the
obvious, given your less than accurate/complete "analysis" of the problem;
clearly, if there was a fundamental backend database problem then why hasn't
it affected my OpenLDAP 1.2.7 test, or my 2.0-devel system (with over 20,000
entries across two database backends)?

>From your original message:

  "I hoped to find a working version in the cvs tree but currently it
  doesn't compile."

Precisely what is "it" (ie., using what CVS tag did you download source)?  My
test system (derived from CVS) compiles and runs fine.