changes for 2.0.24

["Howard Chu" <hyc@highlandsun.com>]

This would be a good starting point.

>>>>
-----Original Message-----
From: owner-openldap-committers@OpenLDAP.org
[mailto:owner-openldap-committers@OpenLDAP.org]On Behalf Of Kurt D.
Zeilenga
Sent: Monday, April 15, 2002 7:33 AM
To: Howard Chu
Cc: openldap-committers@OpenLDAP.org
Subject: Re: bugfix disposition


At 01:38 AM 2002-04-13, Howard Chu wrote:
>Most of the active bug reports have been patched in HEAD, but only a few
>have been fixed in the 2.0 Release stream. Also, there are some pretty old
>bug reports still in "Test" state that should be either Release or Closed
by
>now. Assuming there will be a 2.0.24 I think we ought to take care of some
>others as well.

I assume there will be a .24, .25, ...

I'm fine with bringing small bug fixes into OPENLDAP_REL_ENG_2,
I just don't have time to back port many of the larger fixes.
In some cases, I rather tell folks that the fix is to upgrade.

Kurt

>ITS#818 - the bug was filed against 2.0.6 and was known to be working in
>2.0.18, and is apparently working in HEAD as of Dec 7 2001.
>        back-ldbm/search.c

Large back port job.

hyc< I haven't looked at the extent of these patches yet. I see that 2.0.23
hyc< fails the test with the scripts in ITS#1234 though.

>ITS#910 - filed against 2.0.7 and left Suspended. Seems this code hasn't
>ever changed, but the bug itself is pretty unlikely in the first place.
>        liblber/sockbuf.c

A fix for this is needed.

hyc< This is pretty surprising to me. The submitted patch looks fine, but I
hyc< was unable to reproduce any failure on Linux. Perhaps I need to use a
hyc< larger group, but the one I used was already 26K in size. I suppose an
hyc< entry with a jpeg file or somesuch would do just as well.

>ITS#980 - filed against 2.0.7 but the implication is that the bug
>disappeared after upgrading a server to 2.0.7 - ???

2.0 referral handling is broken, more broken in some versions than
others.  Fixed mostly by ITS#818.

hyc< Right. I later realized the problem had to do with v3 referrals, as
hyc< opposed to v2. The comments about not being able to test after
upgrading
hyc< were in reference to the protocol versions...

>ITS#1022 - filed against 2.0.7, fixed 2001-06-09 in HEAD.
>        maildap/main.c 1.5

This should be applied to 2.0.

hyc< Easy patch, delete one argument to sprintf...
	maildap/main.c 1.4 -> 1.5

>ITS#1234 - filed against 2.0.11, ITS#818 mentions that 1234 is fixed as of
>2001-12-07 in HEAD
>        back-ldbm/search.c 1.68

The ITS implies referral code still needs work.

>ITS#1289 - filed against 2.0.11, fixed 2001-09-08 in HEAD.
>        back-ldbm/search.c 1.61

See #818.

hyc< right, need to examine how much code is involved.

>ITS#1335 - filed against 2.0.19, fixed 2001-09-19 in HEAD, needs a
>decision...

Ando patch likely should be back ported.

>ITS#1508 - filed against 2.0.19, fixed 2002-01-12 in HEAD
>        slapd/acl.c 1.120

I view this as an enhancement to experimental code, so don't
think this should be back ported.

>ITS#1570 - current, patched 2002-04-11 in HEAD
>        back-ldbm/idl.c 1.68

Should be back ported.

hyc< back-ldbm/idl.c 1.67 -> 1.68; I'm fairly sure that only the line
hyc< around 1080 is needed to avoid a memory overwrite. It prevents any
hyc< dups in (a) from propagating. The check at 1067 prevents any dups
hyc< in (b) from propagating. I never actually saw any bad IDLs with dups
hyc< present in my own testing.

>ITS#1613 - current, patched 2002-04-11 in HEAD
>        slurpd/main.c 1.24

Should be back ported.

hyc< Looks like slurpd/main.c 1.23 -> 1.26 for this.

>ITS#1615 - current, fixed 2002-03-07 in HEAD
>        slapd/daemon.c 1.212

Should be back ported.

hyc< slapd/daemon.c 1.211 -> 1.212

>ITS#1649 - current, fixed 2002-03-18 in HEAD
>        tools/ldapsearch.c 1.112

Should be back ported.

hyc< clients/tools/ldapsearch.c 1.111 -> 1.112

>ITS#1655 - current, fixed 2002-03-20 in HEAD
>        slapd/connection.c 1.166

I need to look at this... but likely okay to back port.

>ITS#1660 - current, (partial) fix 2002-03-23 in HEAD
>        slapd/connection.c 1.167

I think the partial fix (the typo) should be minimally back ported.

hyc< slapd/connection.c 1.165 -> 1.166 is straightforward,
hyc< from 1.166 to 1.167 is affected by berval'ization of strings but it's
hyc< not a big patch.

>ITS#1721 - current, fixed 2002-03-10 in HEAD
>        slapd/extended.c 1.28

Should be back ported.
hyc< I don't see a clear set of HEAD revisions that provides this fix.
hyc< it's a small patch to generate though, line 86 needs to add " = NULL;"
hyc< initialization. Line 152, I think needs to check
hyc<   reqdata && reqdata->bv_len ? reqdata : NULL

>ITS#1728 - current, fixed in HEAD
>        slapd/daemon.c 1.204, 1.212

Should be back ported.

hyc< slapd/daemon.c 1.203 -> 1.204; ITS#1615 pulls in daemon.c 1.211 ->
1.212

>The rest of the current bugs are filed against 2.1alpha or HEAD, and many
of
>them are to features that don't exist in 2.0. I would just as soon ignore
>them for 2.0, but I guess ITS#1740 might merit a backport.

I concur.
<<<<

  -- Howard Chu
  Chief Architect, Symas Corp.       Director, Highland Sun
  http://www.symas.com               http://highlandsun.com/hyc
  Symas: Premier OpenSource Development and Support