(ITS#5256) encoding of guide.sdf
by dieter＠dkluenter.de
Full_Name: Dieter Kluenter
Version: REL_ENG_2_4
OS: Linux
URL: ftp://ftp.openldap.org/incoming/
Submission from: (NULL) (84.142.202.173)
Hi,
the guide *.sdf files are UTF-8 encoded, but the online availeable HTML files
seem to be iso8859-1 encoded, which leeds to display errors, an example is part
9.2.1.
... proxy’s efficiency ...
-Dieter
15 years, 9 months
Re: (ITS#5254) core dump on Solaris 9 when with threads
by hyc＠symas.com
mills(a)cc.umanitoba.ca wrote:
> Full_Name: Gary Mills
> Version: openldap-2.3.38
> OS: Solaris 9 SPARC
> URL: ftp://ftp.openldap.org/incoming/
> Submission from: (NULL) (24.78.105.29)
back-ldbm is deprecated and has been deleted from OpenLDAP 2.4. It is known to
have many problems. Berkeley DB 3.1.17 is ancient, and also known to have many
bugs. Don't use ldbm, and use a recent version of BerkeleyDB. Nobody is going
to spend any time investigating these broken code bases.
>
>
> Here's how openldap-2.3.38 was built:
>
> PATH=/opt/SUNWspro/bin:/usr/bin:/usr/ccs/bin:/usr/sbin; export PATH
>
> env CC=cc \
> INSTALL=/usr/ucb/install \
> CPPFLAGS="-I/usr/local/src/db/db-3.1.17/build_unix
> -I/usr/local/src/OpenSSL/openssl-0.9.8f/include" \
> LDFLAGS="-R/usr/local/lib -L/usr/local/src/db/db-3.1.17/build_unix
> -L/usr/local/src/OpenSSL/openssl-0.9.8f/lib" \
> ./configure \
> --disable-proctitle \
> --enable-crypt \
> --disable-bdb \
> --disable-hdb \
> --enable-ldbm \
> --enable-perl
>
> In simple tests with an ldbm database, it works correctly. With many clients,
> it
> dump core hundreds of times a day. The dumps vary, but suggest a memory
> management
> problem. Here are two:
--
-- Howard Chu
Chief Architect, Symas Corp. http://www.symas.com
Director, Highland Sun http://highlandsun.com/hyc/
Chief Architect, OpenLDAP http://www.openldap.org/project/
15 years, 9 months
Re: (ITS#5254) core dump on Solaris 9 when with threads
by mills＠cc.umanitoba.ca
I neglected to mention that compiling Openldap without threads resulted
in a perfectly stable server. Threads seem to be handled incorrectly,
somehow.
--
-Gary Mills- -Unix Support- -U of M Academic Computing and Networking-
15 years, 9 months
(ITS#5255) Outlook 2003 says: The search could not be completed.
by mills＠cc.umanitoba.ca
Full_Name: Gary Mills
Version: openldap-2.3.38
OS: Solaris 9 SPARC
URL: ftp://ftp.openldap.org/incoming/
Submission from: (NULL) (24.78.105.29)
Here's how Openldap was compiled:
PATH=/opt/SUNWspro/bin:/usr/bin:/usr/ccs/bin:/usr/sbin; export PATH
env CC=cc \
INSTALL=/usr/ucb/install \
CPPFLAGS="-I/usr/local/src/db/db-3.1.17/build_unix
-I/usr/local/src/OpenSSL/openssl-0.9.8f/include" \
LDFLAGS="-R/usr/local/lib -L/usr/local/src/db/db-3.1.17/build_unix
-L/usr/local/src/OpenSSL/openssl-0.9.8f/lib" \
./configure \
--disable-proctitle \
--enable-crypt \
--disable-bdb \
--disable-hdb \
--enable-ldbm \
--without-threads \
--enable-perl
When using an ldbm database, Microsoft Outlook 2003 clients report this error
when
they attempt a search:
The search could not be completed. MAPI_E_CALL_FAILED
The LDAP logs show this message:
critical control unavailable in context
This happens because Openldap advertizes the pagedResults control, even though
the
ldbm backend does not support it. When OL 2003 obtains the list of available
controls, the LDAP server does not know which back end it will use. When OL
2003
does a query, it goes to the ldbm back end which does not support that control.
OL 2003 marks it critical, making the situation impossible.
In this situation, I'm using only the ldbm back end, so I removed the
pagedResults
control in servers/slapd/controls.c to fix the problem. Here's my patch:
*** controls.Oc Tue Jan 2 15:43:55 2007
--- controls.c Thu Nov 29 19:53:12 2007
***************
*** 123,132 ****
--- 123,134 ----
SLAP_CTRL_GLOBAL|SLAP_CTRL_SEARCH, NULL,
parseValuesReturnFilter, LDAP_SLIST_ENTRY_INITIALIZER(next) },
#endif
+ #ifdef VANILLA
{ LDAP_CONTROL_PAGEDRESULTS,
(int)offsetof(struct slap_control_ids, sc_pagedResults),
SLAP_CTRL_SEARCH, NULL,
parsePagedResults, LDAP_SLIST_ENTRY_INITIALIZER(next) },
+ #endif
#ifdef LDAP_DEVEL
{ LDAP_CONTROL_SORTREQUEST,
(int)offsetof(struct slap_control_ids, sc_sortedResults),
Adding an access list to the config file did not work.
15 years, 9 months
(ITS#5254) core dump on Solaris 9 when with threads
by mills＠cc.umanitoba.ca
Full_Name: Gary Mills
Version: openldap-2.3.38
OS: Solaris 9 SPARC
URL: ftp://ftp.openldap.org/incoming/
Submission from: (NULL) (24.78.105.29)
Here's how openldap-2.3.38 was built:
PATH=/opt/SUNWspro/bin:/usr/bin:/usr/ccs/bin:/usr/sbin; export PATH
env CC=cc \
INSTALL=/usr/ucb/install \
CPPFLAGS="-I/usr/local/src/db/db-3.1.17/build_unix
-I/usr/local/src/OpenSSL/openssl-0.9.8f/include" \
LDFLAGS="-R/usr/local/lib -L/usr/local/src/db/db-3.1.17/build_unix
-L/usr/local/src/OpenSSL/openssl-0.9.8f/lib" \
./configure \
--disable-proctitle \
--enable-crypt \
--disable-bdb \
--disable-hdb \
--enable-ldbm \
--enable-perl
In simple tests with an ldbm database, it works correctly. With many clients,
it
dump core hundreds of times a day. The dumps vary, but suggest a memory
management
problem. Here are two:
# mdb /usr/local/libexec/slapd /var/core/core.slapd.3610
Loading modules: [ libc.so.1 libthread.so.1 ld.so.1 ]
> ::status
debugging core file of slapd (32-bit) from electra
executable file: /opt/local/libexec/slapd
initial argv: slapd
threading model: multi-threaded
status: process terminated by SIGSEGV (Segmentation Fault)
> ::stack
is_ad_subtype+0x4c(30000, 3b0120, 33abb0, 33a698, 181ae30, 15949a4)
attrs_find+0x2c(181b458, 3b0120, 33a698, 400, 181ae30, 15949a4)
0x8cba4(3d0e08, 17d4e80, 15949c4, 3bd000, 0, fdbff725)
test_filter+0x388(3d0e08, 17d4e80, 15949c4, 14baa8, ffffffe0, fdbff8b1)
0x8c808(3d0e08, 17d4e80, 15949c4, 7, ffffffe0, fdbff8f1)
test_filter+0x784(3d0e08, 17d4e80, 15949d4, 2, 0, 0)
ldbm_back_search+0x12d0(3d0e08, fdbffd50, 0, fdbff2a4, 0, fdbff9fd)
fe_op_search+0x594(3d0e08, fdbffd50, 3d0efc, fdbffab0, 0, 0)
do_search+0xf54(3d0e08, fdbffd50, fdbffd40, 0, 0, 0)
0x5bbf8(fdbffe14, 3d0e08, 0, 0, 0, 0)
0x153f5c(3aacb8, 0, 0, 0, 0, 0)
libthread.so.1`_lwp_start(0, 0, 0, 0, 0, 0)
# mdb /usr/local/libexec/slapd /var/core/core.slapd.25318
Loading modules: [ libc.so.1 libthread.so.1 ld.so.1 ]
> ::status
debugging core file of slapd (32-bit) from electra
executable file: /opt/local/libexec/slapd
initial argv: slapd
threading model: multi-threaded
status: process terminated by SIGSEGV (Segmentation Fault)
> ::stack
str2entry2+0xfec(40c010, 0, 2ee1bc, 7, ffffffe0, fe3ff8f1)
id2entry_rw+0x484(3d0820, 1da9, 0, 2, 0, 0)
ldbm_back_search+0xa14(3d0e08, fe3ffd50, 0, fe3ff2a4, 81010100, ff00)
fe_op_search+0x594(3d0e08, fe3ffd50, 3d0efc, 14388f4, fe3ffab0, 1438e4c)
do_search+0xf54(3d0e08, fe3ffd50, fe3ffd40, 0, 0, 0)
0x5bbf8(fe3ffe14, 3d0e08, 0, 0, 0, 0)
0x153f5c(3aacb8, 0, 0, 0, 0, 0)
libthread.so.1`_lwp_start(0, 0, 0, 0, 0, 0)
15 years, 9 months
Re: (ITS#5252) Several SQL Bugs in back-sql
by ando＠sys-net.it
h.goebel(a)goebel-consult.de wrote:
> Full_Name:
> Version: 2.3.38
> OS:
> URL: ftp://ftp.openldap.org/incoming/
> Submission from: (NULL) (79.207.201.247)
>
>
> Debugging marks in sql-statement may lead to wrong results.
>
> The interface I implemented yields statements like this when using Thunderbird
> for searching for 'Klaus'. Thunderbird tries to look up in email, name and given
> name.
> Ans in my setup, search is restricted to "(objectclass=evolutionPerson)".
>
> SELECT DISTINCT
> ldap_entries.id,res_partner_address.id,
> text('evolutionPerson') AS objectClass,ldap_entries.dn AS dn FROM
> ldap_entries,res_partner_address,ldap_entry_objclasses
> WHERE
> res_partner_address.id=ldap_entries.keyval
> AND ldap_entries.oc_map_id=1
> AND 9=9
> AND ((2=2 OR ( ldap_entries.id=ldap_entry_objclasses.entry_id
> AND ldap_entry_objclasses.oc_name='evolutionPerson'))
> AND (( upper(email) LIKE '%KLAUS%')
> OR (upper(name) LIKE '%KLAUS%')
> OR 7=7
> OR (upper(substring(name from '\\S+..$')) LIKE '%KLAUS%')))
>
> 1) "OR 7=7 OR" renders the whole last AND-clause useless: No selection is done
> on email, name or givenname. Thus all records are returned.
>
> 2) "2=2 OR" renders the restriction to a specific objectclass useless.
>
> In Bug 5198 Pierangelo Masarati say, that a good RDBMs is optimizing this
> part away. And exactly this is the problem.
This is required to implement attribute and objectClass inheritance,
which is required by LDAP. So please don't call them "bugs", they're
required hacks. If they're not in place, back-sql would give the wrong
response to queries; if they're in place, back-sql would give the right
response, possibly inefficiently on braindead RDBMSes. Please choose
what you prefer: the wrong answer quickly or the right one maybe slower?
p.
Ing. Pierangelo Masarati
OpenLDAP Core Team
SysNet s.r.l.
via Dossi, 8 - 27100 Pavia - ITALIA
http://www.sys-net.it
---------------------------------------
Office: +39 02 23998309
Mobile: +39 333 4963172
Email: pierangelo.masarati(a)sys-net.it
---------------------------------------
15 years, 9 months
