openldap-bugs February 2013

openldap-bugs@openldap.org

22 participants
48 discussions

(ITS#7533) Don't install DB_CONFIG.example if built w/o back-hdb or back-bdb
by quanah＠OpenLDAP.org 27 Feb '13

27 Feb '13

Full_Name: Quanah Gibson-Mount Version: 2.4.33 OS: Linux 2.6 URL: ftp://ftp.openldap.org/incoming/ Submission from: (NULL) (74.196.25.250) When building without BDB support, DB_CONFIG.example is still copied to the state directory. Given a lack of BDB support, this seems pointless. installing slapd.ldif in /opt/zimbra/openldap-2.4.33.4z/etc/openldap ../../build/shtool install -c -m 600 slapd.ldif.tmp /opt/zimbra/openldap-2.4.33.4z/etc/openldap/slapd.ldif ../../build/shtool install -c -m 600 ./DB_CONFIG \ /opt/zimbra/openldap-2.4.33.4z/var/openldap-data/DB_CONFIG.example ../../build/shtool install -c -m 600 ./DB_CONFIG \ /opt/zimbra/openldap-2.4.33.4z/etc/openldap/DB_CONFIG.example ../../build/shtool mkdir -p /opt/zimbra/openldap-2.4.33.4z/etc/openldap/schema

1 0

(ITS#7532) File descriptor needed before sending a request
by npmccallum＠redhat.com 27 Feb '13

27 Feb '13

Full_Name: Nathaniel McCallum Version: git master OS: fedora rawhide URL: https://github.com/npmccallum/openldap/commit/4e12a8ca31bcc6ad91f036b9200d7… Submission from: (NULL) (76.177.249.65) For asynchronous usage, it is common to create a "write" callback and perform all request sending in this callback from an internal queue. However, since ldap_open() was deprecated, there is no way to ensure a connection is established before the request is sent. This patch adds a function: int ldap_connect(LDAP *ld). You can call this function immediately after ldap_initialize() to ensure that a file descriptor is created for polling. No default behaviour changes, so applications that don't call ldap_connect() will always get the default lazy connection. The attached file is derived from OpenLDAP Software. All of the modifications to OpenLDAP Software represented in the following patch(es) were developed by Red Hat. Red Hat has not assigned rights and/or interest in this work to any party. I, Nathaniel McCallum am authorized by Red Hat, my employer, to release this work under the following terms. Red Hat hereby places the following modifications to OpenLDAP Software (and only these modifications) into the public domain. Hence, these modifications may be freely used and/or redistributed for any purpose with or without attribution and/or other notice.

1 0

Re: (ITS#7515) Nested liblmdb transaction bugs
by h.b.furuseth＠usit.uio.no 27 Feb '13

27 Feb '13

MDB can act on P_DIRTY without checking whether the page is dirty in this txn's dirty_list, or only some ancestor txn's: * mdb_cursor_put() line 5094, overwriting an F_BIGDATA node: Let a nested txn replace a big item whose page is dirty only in an ancestor txn. The parent will see the new value if the old page had room for it, even if the child aborted. Test program enclosed. Proposed fix: No need to search dirty_list, mdb_page_get() just did. Make it return where the page was found instead of an error code: <0 not found, 0 in map, >0 txn nesting level of the dirty_list with the page. Branch mdb/its7515 in <http://folk.uio.no/hbf/OpenLDAP/openldap.git>, except the commit message may need tweaks. * mdb_xcursor_init1(): Sets DB_DIRTY if mc_top has P_DIRTY. Seems to cancel some MDB_PS_MODIFY activity. Is that OK? #include <lmdb.h> #include <stdio.h> int main(void) { static char pTxt[9999] = "parent", cTxt[9999] = "child"; MDB_val key = {1,""}, pVal = {9999,pTxt}, cVal = {9999,cTxt}, val; MDB_env *env; MDB_txn *txn, *ctxn; MDB_dbi dbi; mdb_env_create(&env); mdb_env_open(env, "test.mdb", MDB_NOSUBDIR, 0666); mdb_txn_begin(env, 0, 0, &txn); mdb_dbi_open(txn, 0, 0, &dbi); mdb_put(txn, dbi, &key, &pVal, 0); /* put "parent" */ mdb_txn_begin(env, txn, 0, &ctxn); mdb_put(ctxn, dbi, &key, &cVal, 0); /* overwrite with "child" */ mdb_txn_abort(ctxn); mdb_get(txn, dbi, &key, &val); /* should get "parent" */ puts(val.mv_data); /* ...but got "child" */ mdb_txn_commit(txn); mdb_env_close(env); return 0; } -- Hallvard

1 0

Re: (ITS#7531) Syncrepl push delete operation does not recover when slave is unavailable
by jsynacek＠redhat.com 27 Feb '13

27 Feb '13

On 02/27/2013 12:25 PM, jsynacek(a)redhat.com wrote: > Full_Name: Jan Synacek > Version: RE24 (7b03a58) > OS: Linux - Fedora 18 > URL: ftp://ftp.openldap.org/incoming/jsynacek-reproducer.tar.gz > Submission from: (NULL) (209.132.186.34) > > > The setup is a master with an ldap backend and a slave consumer. If the master > is running and slave is not, delete operation results in the master being > stuck. > > Steps to reproduce: > 1. Start master. > 2. Add the example ldif. > 3. Delete one of the test org units. > 4. Observe the error. > > I would expect the master to honor the retry setting. > During the reproduction, it may be necessary to first have both master and slave running, delete a testing test org unit (so the first replication is successful), stop the slave and then continue to step 3. -- Jan Synacek Software Engineer, BaseOS team Brno, Red Hat

1 0

(ITS#7531) Syncrepl push delete operation does not recover when slave is unavailable
by jsynacek＠redhat.com 27 Feb '13

27 Feb '13

Full_Name: Jan Synacek Version: RE24 (7b03a58) OS: Linux - Fedora 18 URL: ftp://ftp.openldap.org/incoming/jsynacek-reproducer.tar.gz Submission from: (NULL) (209.132.186.34) The setup is a master with an ldap backend and a slave consumer. If the master is running and slave is not, delete operation results in the master being stuck. Steps to reproduce: 1. Start master. 2. Add the example ldif. 3. Delete one of the test org units. 4. Observe the error. I would expect the master to honor the retry setting.

1 0

(ITS#7530) gethostbyname_r fails if initial buffer size is too small
by ujvari＠microsec.hu 26 Feb '13

26 Feb '13

Full_Name: Aron Ujvari Version: 2.4.33 OS: Linux URL: ftp://ftp.openldap.org/incoming/ Submission from: (NULL) (213.163.11.62) OpenLDAP 2.4.33 (and before) does not handle properly when in function ldap_pvt_gethostbyname_a the 6-args version of gethostname_r returns ERANGE(34) error code because of too small buffer size. Although there is a loop in the function to increment the buffer size, but there is no check for ERANGE(34) result in the "r" variable, so no use of the loop. Relevant code is in libraries/libldap/util-int.c, lines 336-368. The gethostbyname man page clearly states at GNU Extensions section that gethostbyname_r can return non-zero value on failure, and return ERANGE if "buf" is too small, and the call should be retried with a larger buffer in this case.

1 0

Re: (ITS#7517) mdb_dbi_close(dbi updated in existing txn) breaks
by h.b.furuseth＠usit.uio.no 19 Feb '13

19 Feb '13

> If a thread modifies a named MDB database and closes it before > committing, the changes are lost. Documented "don't do that". mdb_dbi_close() could defer fully closing the DBI to the current write transaction if one exists. But that would require a mutex locked by [I think] mdb_env_commit(write txn), mdb_dbi_open, and mdb_dbi_close. > Test case (no nested txns): Bogus test program. The mdb_dbi_open() should use a named database: > E(mdb_dbi_open(txn, NULL, rdonly ? 0 : MDB_CREATE, &dbi)); E(mdb_dbi_open(txn, "db", rdonly ? 0 : MDB_CREATE, &dbi)); -- Hallvard

1 0

Re: (ITS#7529) mdb.c missing #if MDB_DEBUG at line 3514
by mark＠ibiblio.org 19 Feb '13

19 Feb '13

Thanks for the follow up Howard. I'm busy cranking out a number of dependent packages to compile some software, OpenLDAP being one of many. We have an aversion to the use of GCC here so situations like this compilation fail are somewhat common. It's nice though when code is kept simple and portable as it makes life easier. I don't normally spend much time debugging code and shipping fixes in a nice box with a pretty bow on it, I've just cut and pasted the two lines I used to allow compilation, part of the compile recipe used to crank out this 2.4.33 release. Hopefully that's enough to update your git archives. ta, Mark. On Wed, Feb 20, 2013 at 3:20 PM, Howard Chu <hyc(a)symas.com> wrote: > mark(a)ibiblio.org wrote: >> >> Full_Name: Mark >> Version: 2.4.33 >> OS: Solaris 10 & 11 >> URL: ftp://ftp.openldap.org/incoming/ >> Submission from: (NULL) (203.63.89.15) >> >> >> The stock 2.4.33 fails to compile mdb with this error: >> >> cc -I/usr/local/include -L/usr/local/lib -I../../../include >> -I../../../include >> -I.. -I./.. -I./../../../libraries/libmdb -I/usr/local/include >> -L/usr/local/lib >> -c ./../../../libraries/libmdb/mdb.c -o mdb.o >> "./../../../libraries/libmdb/mdb.c", line 3514: undefined symbol: top >> "./../../../libraries/libmdb/mdb.c", line 3514: left operand of "->" must >> be >> pointer to struct/union >> cc: acomp failed for ./../../../libraries/libmdb/mdb.c >> >> The problem is top is defined inside #if MDB_DEBUG but used without the >> same #if >> qualifier. > > > Normally the DPRINTF macro would just expand to ((void) 0) if MDB_DEBUG is > not defined, so no, top is not being used without MDB_DEBUG. But looking > closer, I see this requires a C preprocessor that supports vararg macros. > Don't you have a C99-compatible toolchain on Solaris? > >> The fix is: >> /bin/perl -pe '($. == 3514) && s%^%#if MDB_DEBUG\n%' -i >> libraries/libmdb/mdb.c >> /bin/perl -pe '($. == 3517) && s%^%#endif\n%' -i >> libraries/libmdb/mdb.c > > > In the future, please use unified diff format. Preferably just use "git > format-patch" like it says on > http://www.openldap.org/devel/contributing.html > > We don't make changes in already released code, and line numbers alone are a > poor reference in development code. > > -- > -- Howard Chu > CTO, Symas Corp. http://www.symas.com > Director, Highland Sun http://highlandsun.com/hyc/ > Chief Architect, OpenLDAP http://www.openldap.org/project/

1 0

Re: (ITS#7529) mdb.c missing #if MDB_DEBUG at line 3514
by hyc＠symas.com 19 Feb '13

19 Feb '13

mark(a)ibiblio.org wrote: > Full_Name: Mark > Version: 2.4.33 > OS: Solaris 10 & 11 > URL: ftp://ftp.openldap.org/incoming/ > Submission from: (NULL) (203.63.89.15) > > > The stock 2.4.33 fails to compile mdb with this error: Fixed in git rev 0108327c274a4b260b5b355d2a7c3ffc3a85f4c1 2012-10-25, closing this ITS. > > cc -I/usr/local/include -L/usr/local/lib -I../../../include -I../../../include > -I.. -I./.. -I./../../../libraries/libmdb -I/usr/local/include -L/usr/local/lib > -c ./../../../libraries/libmdb/mdb.c -o mdb.o > "./../../../libraries/libmdb/mdb.c", line 3514: undefined symbol: top > "./../../../libraries/libmdb/mdb.c", line 3514: left operand of "->" must be > pointer to struct/union > cc: acomp failed for ./../../../libraries/libmdb/mdb.c > > The problem is top is defined inside #if MDB_DEBUG but used without the same #if > qualifier. > > The fix is: > /bin/perl -pe '($. == 3514) && s%^%#if MDB_DEBUG\n%' -i > libraries/libmdb/mdb.c > /bin/perl -pe '($. == 3517) && s%^%#endif\n%' -i libraries/libmdb/mdb.c > > -- -- Howard Chu CTO, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/

1 0

Re: (ITS#7529) mdb.c missing #if MDB_DEBUG at line 3514
by hyc＠symas.com 19 Feb '13

19 Feb '13

mark(a)ibiblio.org wrote: > Full_Name: Mark > Version: 2.4.33 > OS: Solaris 10 & 11 > URL: ftp://ftp.openldap.org/incoming/ > Submission from: (NULL) (203.63.89.15) > > > The stock 2.4.33 fails to compile mdb with this error: > > cc -I/usr/local/include -L/usr/local/lib -I../../../include -I../../../include > -I.. -I./.. -I./../../../libraries/libmdb -I/usr/local/include -L/usr/local/lib > -c ./../../../libraries/libmdb/mdb.c -o mdb.o > "./../../../libraries/libmdb/mdb.c", line 3514: undefined symbol: top > "./../../../libraries/libmdb/mdb.c", line 3514: left operand of "->" must be > pointer to struct/union > cc: acomp failed for ./../../../libraries/libmdb/mdb.c > > The problem is top is defined inside #if MDB_DEBUG but used without the same #if > qualifier. Normally the DPRINTF macro would just expand to ((void) 0) if MDB_DEBUG is not defined, so no, top is not being used without MDB_DEBUG. But looking closer, I see this requires a C preprocessor that supports vararg macros. Don't you have a C99-compatible toolchain on Solaris? > The fix is: > /bin/perl -pe '($. == 3514) && s%^%#if MDB_DEBUG\n%' -i > libraries/libmdb/mdb.c > /bin/perl -pe '($. == 3517) && s%^%#endif\n%' -i libraries/libmdb/mdb.c In the future, please use unified diff format. Preferably just use "git format-patch" like it says on http://www.openldap.org/devel/contributing.html We don't make changes in already released code, and line numbers alone are a poor reference in development code. -- -- Howard Chu CTO, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/

1 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

openldap-bugs February 2013